Subject: Space-tech Digest #117 Contents: Magnetic sails (12 msgs) ------------------------------------------------------------ Date: Wed, 22 Apr 1992 02:41 EDT From: "GORDON D. PUSCH" To: space-tech@cs.cmu.edu, Bradford@maccs.McMaster.CA, brand <@cunyvm.cuny.edu:brand@vtcc1.BITNET> Subject: Magnetic sails (long) There's a cover-story article I've just read that I think a lot of you will be interested in: ``The Magnetic Sail,'' by Robert M. Zubrin (based on work by Dana G. Andrews (Boeing) and himself (Martin-Marietta). I regret it appeared in an only *quasi*-reliable source: the "science fact" section of this month's (May 1992) _Analog_. I haven't seen the original references, 'cause we ain't got'em here, and I don't have time right now to attempt to rederive his formulas, 'cause I'm on a tight deadline this week ... ('though I can *almost* see what approximations they used by eyeball'n'em). The gist of their idea is: use the magnetic field of a current loop as a ``sail'' for the solar wind or Geomagnetic field to ``push'' against. They calculate the plasma drag of the (1 AU-radius) solar wind against the ``artificial magnetosphere'' of a 200km-circumference loop of super- conducting wire is an unbelievable *86 Newtons*! Their fiducial design masses 5 tonnes, and it can haul itself plus a 14 tonne payload to Mars (and protect its payload from solar flares) in no more than a mere 283 days! It can go from LEO to Earth-escape by ``pumping'' against the Earth's dipole field in only 76.6 days! It can decelerate (while protecting the payload of) an interstellar probe from 0.1c to planetary velocities in a mere 55.5 years! It might even be capable of lifting off from the *Earth's surface* (but only from Bathurst island, or the Adelie coast of Antarctica ;-) ... And all of these wonderful good things using *NO FUEL* ... 8-* (see P.S.) The large loop size (200km circumferance) is driven by the desire to minimize mass while maximizing drag; they find this requires them (for a given material) to maximize sail-radius and minimize current, as shown by their formula for the Drag/Mass ratio: D/M = (0.6) * (J_c/rho_m) * ( mu_0 * R_m * (rho_s * V_s^2)^2 / I )^(1/3), where: J_c --- critical current-density ( 1*10^10 Amp/m^2 for NbTi ), rho_m --- magsail superconductor's mass-density ( 5000 kg/m^3 ), I --- magsail's total current ( ~50 kAmp, they figure ) R_m --- magsail's radius ( ~32 km, they figure ), rho_s --- solar-wind mass-density ( 8.31*10^-21 kg/m^3 @ 1 AU ) V_s --- solar-wind velocity ( 400--600 km/sec @ 1AU ) mu_0 --- ``permiability of free space'' ( 4*pi * 10^-7 N/Amp^2 ). The magnetic at the loop center is small --- about 10 microTesla --- and the stored energy in the field is about 80 MJ; a 10KW generator could ``inflate'' the sail in about 2.2 hours. He says they assume ``not unreasonable'' (my quotes) properties for their superconductor: i.e., density comparable to the CuO high-T_c ceramics, but J_c and tensile strength comparable to Nb_3-Sn or Nb-Ti. They assume high-T_c's because they claim that LN2-temperatures can be easily obtained and maintained by radiative cooling at 1 AU, but LHe can't --- and that even if the tensile strength were *zero*, reinforcing it with aluminum would only add about 10% to its mass (he doesn't mention it, but even NbTi has to be imbedded in a Cu or Al matrix to thermally stabilize it and protect it against quenches, so the high-T_c's will probably also need a matrix anyway). The magsail's terminal velocity is on the order of the solar wind's: 400--600 km/sec. Because its drag can easily exceed its weight, like a solar sail, it can easily ``loiter'' in an orbit at less than circular velocity, waiting for its target to ``catch up;'' it therefore cares not a whit about such frumpery as ``launch windows.'' By making the (IMHO questionable) assumption that the surface where the solar-wind stagnation-pressure and magnetic ``pressure'' are equal (definition of ``magnetopause'' surface?) could be plugged into a hyper- sonic aerodynamics code, they find that, depending on the ``angle of attack'' (angle of the loop-plane to the solar wind) it can develop ``lift'' as well as drag (L/D of about 0.14 for a simple dipole --- not much worse than the shuttle :-). Although ``wants'' to align its magnetic moment w/ the ambient magnetic field, it can maintain itself at an angle to the wind by shifting its payload mass off-center, to use the inertial ``torque'' caused by its acceleration to balance the magnetic torque; I suspect this will require *actively* controlling the payload-mass's position --- past some angle it'll be in an unstable equilibrium. Assuming ``not unreasonable'' advances in superconductor technology (factor of 80 increase in J_c/{mass-density} over NbTi (!)), and that the loop can be stabilized with its magnetic moment pointing the ``wrong way 'round'' to repel the Earth's moment (again, an unstable equilibrium; it *wants* to flip over, as anyone who's ever played with a magnet can verify) it can lift itself plus 4 tonnes at 3.27 gees (from bathurst island! :-), with acceleration dropping like r^-4 (dipole-dipole interaction!). Treating dipole-dipole force as a potential, and integrating to ``infinity,'' they estimate a final velocity of 11.68 km/sec, cutting transit-time to Mars to about 150 days ... My gut-feeling physical-intuition hunch is: it all sounds *far* too good to be true. If it *IS* true, though, I can see all *sorts* of interesting possibilities opening up --- including manned exploration of Jovian space. It also occurs to me that the magsail and lightsail complement each other nicely --- magsails can escape from LEO (or maybe even BGZ --- Bathurst Ground Zero ;-), and protect their payloads from flares and van Allens; lightsails probably have more thrust in the inner system (the solar *photon* ``wind'' carries about 100 times the momentum-density of the solar *plasma* wind) --- although perhaps the magsail's mass can be *much* lower for a given thrust, because it's a *loop*, instead of an *area*, and its effective closs-section can be larger than its physical cross-section ... If you *did* decided to combine them, you could use the shrouds connecting the payload to the loop to rig the lightsails from, and hold them rigid ... Any comments? Gordon D. Pusch P.S. --- Except for the ``no fuel'' part, one of these things *sounds* like it would handle a lot like one of A.B. Chandler's ``Gaussjammers'' --- except that it doesn't look ``like a peg-shaped top,'' or have FTL capability, or require a ``diesel emergency-jennie, to run her Erhrenhaft jennies --- since often --- *far* too often --- she'll blunder into a magnetic storm, which flings her to hell and gone, *and* drains her pile to lead ...'' not that there's anywhere to *go*, mind you, since ``... she's off the beaten tramlines, hopelessly lost, and must fly on and on and on, like some sort of latter-day flying dutchman, with her stinking diesel fouling her air, and her whole crew starving --- until they stumble across some habitable world, to found yet *another* `lost colony' for the novelists to write about ...'' [rough quote from memory] ;-) ------------------------------ Date: Wed, 22 Apr 92 07:56:54 EDT From: John Roberts Disclaimer: Opinions expressed are those of the sender and do not reflect NIST policy or agreement. To: space-tech@cs.cmu.edu Subject: Re: Magnetic sails >They assume high-T_c's because >they claim that LN2-temperatures can be easily obtained and maintained by >radiative cooling at 1 AU, but LHe can't --- and that even if the tensile >strength were *zero*, reinforcing it with aluminum would only add about 10% >to its mass (he doesn't mention it, but even NbTi has to be imbedded in a >Cu or Al matrix to thermally stabilize it and protect it against quenches, >so the high-T_c's will probably also need a matrix anyway). According to Larry Niven (and a few more reputable sources :-), one characteristic of superconductors is that they maintain the same temperature throughout, and in fact temperature changes propagate at a rate not too much slower than the speed of light. If this is true, then you might only need to apply active cooling at one point in the loop. John Roberts roberts@cmr.ncsl.nist.gov ------------------------------ Date: Wed, 22 Apr 92 09:29:52 -0400 From: dietz@cs.rochester.edu To: PUSCHG@crl.aecl.ca Subject: Re: Magnetic sails (long) Cc: space-tech@cs.cmu.edu One problem they glossed over in the "launch the magsail from earth" scenario is that the back-EMF in the loop requires the presence of a large power supply (the loop, moving upwards, is threaded by a changing magnetic flux, so a voltage is induced). For a 5 ton vehicle accelerating at 3 gees at 7 km/s (say), the required power is about 1 gigawatt. Another problem they ignored is the gross inhomogeneity of the solar wind. The solar wind speed varies by almost an order of magnitude, depending on the occurence of flares, coronal holes, etc. Paul F. Dietz dietz@cs.rochester.edu ------------------------------ From: henry@zoo.toronto.edu Date: Wed, 22 Apr 92 11:19:25 EDT Subject: Re: Magnetic sails (long) To: space-tech@cs.cmu.edu >... They assume high-T_c's because >they claim that LN2-temperatures can be easily obtained and maintained by >radiative cooling at 1 AU, but LHe can't... Basically correct. Equilibrium temperature for an object that is well shielded from the Sun and any nearby warm objects, but exposed to space, is circa LN2 temperature at 1 AU. (In space, LOX is a storable propellant!) Getting down much colder is harder; in particular, the average temperature of the sky is something like 25K, so LHe requires active refrigeration for long-term storage. (The cosmic background is circa LHe temperature, but there are stars and such in the sky too, and they warm things up.) Henry Spencer at U of Toronto Zoology henry@zoo.toronto.edu utzoo!henry ------------------------------ From: Christopher Neufeld To: space-tech@cs.cmu.edu Subject: Re: Magnetic sails Date: Wed, 22 Apr 1992 11:26:02 -0400 >According to Larry Niven (and a few more reputable sources :-), one >characteristic of superconductors is that they maintain the same temperature >throughout, and in fact temperature changes propagate at a rate not too much >slower than the speed of light. If this is true, then you might only need >to apply active cooling at one point in the loop. Larry Niven and a few more reputable sources are wrong, in fact almost backwards. This is probably an attempt to fit the semi-classical relation between electrical and thermal conductivity, whcih is more or less a fixed ratio for many fixed metals, in the Sommerfeld model of a metal (a box of non-interacting fermions). This is not a close approximation to a superconductor. There exists in low temperature physics a thermal 'switch', a device which can connect two objects thermally, and disconnect them, without moving parts (which complicate cryostat design, and dump heat where you least want it). You put your sample and your heat source/sink at opposite ends of a piece of superconductor below its critical temperature. By applying a magnetic field you drive it normal, or let it remain superconducting by removing the field. In the normal state the ends are in good thermal contact. In the superconducting state it acts as a reasonably good thermal insulator. In this way it is pretty well exactly opposite to the _Ringworld Engineers_ superconductors. Reference available on request (it's on my desk, but I'm not there, we don't have terminals within eight floors of my office). -- Christopher Neufeld....Just a graduate student | alien: n. a being who neufeld@helios.physics.utoronto.ca Ad astra | travels great distances cneufeld@terranet.cts.com | to molest our cattle "Don't edit reality for the sake of simplicity" | and trample our grain ------------------------------ From: henry@zoo.toronto.edu Date: Wed, 22 Apr 92 11:23:02 EDT Subject: Re: Magnetic sails (long) To: space-tech@cs.cmu.edu >... Equilibrium temperature for an object that is well >shielded from the Sun and any nearby warm objects, but exposed to space, >is circa LN2 temperature at 1 AU... Oops, I meant to add: of course, that shielding is not something you can take for granted. Particularly if you're in orbit around Earth, and need to maintain a specific range of attitudes for some reason (e.g. to keep your magsail sailing), keeping shielding between you and both Earth and Sun could be quite a trick. Existing LN2-temperature space hardware (infrared sensors, some kinds of high-energy spectrometers) tends to use active refrigeration to avoid having to constantly juggle thermal shielding around. Henry Spencer at U of Toronto Zoology henry@zoo.toronto.edu utzoo!henry ------------------------------ Date: Wed, 22 Apr 92 12:57:15 PDT From: gwh@lurnix.COM (George William Herbert) To: space-tech@cs.cmu.edu Subject: Re: Magnetic Sails Bob Zubrin didn't forget about the power supply problem and the inhomogenity of the solar wind; he wrote a "scaled down" version of a tech paper for Analog, and cut out confusing sections and did a bit of speculation. The effects of the variable solar wind on such a system will all in all probably earn someone a PhD somewhere 8-), thus are inappropriate in a "popular" article. And power supply problems for surface launch are obvious to anyone who knows state of the art. He pointed out that the magnets had to do an order of magnitude better, but didn't bother to mention that the power plant had to also. -george ------------------------------ Date: Wed, 22 Apr 1992 17:25 EDT From: "GORDON D. PUSCH" Subject: Re: Magnetic Sails To: space-tech@cs.cmu.edu X-VMS-To: SPACE-TECH X-VMS-Cc: PUSCHG From: Paul F. Dietz : > > One problem they glossed over in the "launch the magsail from earth" > scenario is that the back-EMF in the loop requires the presence of > a large power supply (the loop, moving upwards, is threaded by a > changing magnetic flux, so a voltage is induced). For a 5 ton vehicle > accelerating at 3 gees at 7 km/s (say), the required power is about > 1 gigawatt. > I disagree. Recall that one of the properties of a superconducting loop is that (modulo flux-vortex tunnelling) the flux threading the loop *CAN'T CHANGE*; because it has ``zero'' resistance, the loop-integral of \vec{E} around an unbroken superconducting must be zero, which by the third Maxwell equation, $\curl{\vec{E}} = -\partial\vec{B}/partial{t}$ (sorry about the TeX notation, but I'm tired of trying to remember how ``runoff'' does it) implies that $\int \vec{B} \cdot \vec{dS}$ is *constant*. What *actually* happens is, the ``freezing'' of the back-EMF at zero induces a change in the current, decreasing it to keep the flux *constant*, (again modulo flux-vortex tunnelling); since $V \equiv 0$, the *electrical power* $P = V I \equiv 0$ also. So where is the energy for the launch coming from, you ask? It's being extracted from the *stored energy in the field*, via the magnetic dipole- dipole interaction. There's plenty of energy available --- they assumed improvement by a factor of 80 in $J_c / \rho_m$; assuming the entire increase to be in the current, $J_c$, the stored energy goes up by $80^2$, to 512 GJ --- comparable to the 690 GJ required for a 5 tonne vehicle to escape with $v_final \simeq 11 kps$. I haven't worked out the consequences of decreasing $\rho_m$ yet, but I suspect they'll be similar. Whether that much improvement in $J_c/\rho_m$ is *possible* is a question I'm not competent to answer; any materials-scientists in the house? > > Another problem they ignored is the gross inhomogeneity of the solar > wind. The solar wind speed varies by almost an order of magnitude, > depending on the occurence of flares, coronal holes, etc. > Yeah? So what? All ``windjammer'' skippers had to contend with analogous problems --- it just meant that their ETA's had big error-bars. A ``gauss- jammer'' skipper will have to factor in the probability of getting ``becalmed in the coronal doldrums'' into his/her voyage-plan, and arrange to arrive a bit earlier. Remember, neither light- nor mag-sails give two phigs about launch windows --- if they arrive a week early, it just means they'll need to ``loiter'' a week until Mars catches up to them ... From: Henry Spencer >> ... They assume high-T_c's because they claim that LN2-temperatures >> can be easily obtained and maintained by radiative cooling at 1 AU, >> but LHe can't... > > Basically correct. Equilibrium temperature for an object that is well > shielded from the Sun and any nearby warm objects, but exposed to space, > is circa LN2 temperature at 1 AU. ... the average temperature of the sky > is something like 25K, so LHe requires active refrigeration for long-term > storage ... > ... shielding is not something you can take for granted ... if you're > in orbit around Earth, and need to maintain a specific range of attitudes > (e.g. to keep your magsail sailing), keeping shielding between you and both > Earth and Sun could be quite a trick. Existing LN2-temperature space > hardware ... tends to use active refrigeration to avoid having to > constantly juggle thermal shielding around. > I believe Zubrin described the superconducting loop as sheathed in super- insulation, etc.; while I don't recall him explicitly *saying* so, I suspect from his description that in a detailed *technical* article (which the _Analog_ article *wasn't*, as I see that George Herbert has just pointed out :-), he would have postulated a *minimal* active refrigerating system to avoid exactly the hassle you describe. Since heat pumps are quite efficient when T_lo is nearly equal to T_hi, it'd take relatively little power to actively refrigerate to 25K, but *lots* to refrigerate to 4.5K ... From: Christopher Neufeld >On Apr 22, 7:56am, John Roberts wrote: >> >>According to Larry Niven (and a few more reputable sources :-), >>one characteristic of superconductors is that they maintain the same >>temperature throughout, and in fact temperature changes propagate at >>a rate not too much slower than the speed of light. If this is true, >>then you might only need to apply active cooling at one point in the >loop. > > Larry Niven and a few more reputable sources are wrong, in fact > almost backwards. This is probably an attempt to fit the semi-classical > relation between electrical and thermal conductivity, > whcih is more or less a fixed ratio for many fixed metals, in the > Sommerfeld model of a metal (a box of non-interacting fermions). This is > not a close approximation to a superconductor. > Indeed; the Cooper-pairs behave like a *Bose liquid*, not a Fermi gas ... > There exists in low temperature physics a thermal 'switch', a device > which can connect two objects thermally, and disconnect them, without > moving parts ... You put your sample and your heat source/sink at > opposite >ends of a piece of superconductor below its critical temperature. > By applying a magnetic field you drive it normal, or let it remain > superconducting by removing the field. In the normal state the ends are > in good thermal contact. In the superconducting state it acts as a > reasonably good thermal insulator ... > I'm afraid I don't understand that at *all*. While I agree that Niven's wrong, I thought that heat propagrated through a superconductor the same way it does through any *other* superfluid: by ballistic phonon transport (so-called ``second sound''). How the heck does this thingy work? What's the physics behind it? > Reference available on request (it's on my desk, but I'm not there, > we don't have terminals within eight floors of my office). > Yes, pleasepleaeplease send it! Gordon D. Pusch ------------------------------ Date: Wed, 22 Apr 92 19:28:31 -0400 From: dietz@cs.rochester.edu To: PUSCHG@crl.aecl.ca Subject: Re: Magnetic Sails Cc: space-tech@cs.cmu.edu (I'm sending this again because I forgot to send it to space-tech.) > So where is the energy for the launch coming from, you ask? It's being > extracted from the *stored energy in the field*, via the magnetic > dipole- dipole interaction. There's plenty of energy available --- > they assumed improvement by a factor of 80 in $J_c / \rho_m$; assuming > the entire increase to be in the current, $J_c$, the stored energy > goes up by $80^2$, to 512 GJ --- comparable to the 690 GJ required for > a 5 tonne vehicle to escape with $v_final \simeq 11 kps$. I haven't > worked out the consequences of decreasing $\rho_m$ yet, but I suspect > they'll be similar. > > Whether that much improvement in $J_c/\rho_m$ is *possible* is a question > I'm not competent to answer; any materials-scientists in the house? It would be very hard to do. The virial theorem states that the volume of material needed to confine a system storing E units of magnetic energy (or kinetic energy), goes as E/S, where S is the tensile strength of the material. Put another way, notice that specific energy (energy/mass) and strength/density both have the same units. A system storing enough energy to accelerate itself to 11 km/s has some 60 MJ/kg (you dropped a factor of 2). A material with this tensile strength/density ratio could be formed into a cable 6000 km long capable of supporting itself against 1 gravity. This would be strong enough to build an elevator from the equator to GEO with only a modest taper. If the material has a density of 2, and a modest safety factor, the tensile strength is about 20 million psi. Similar reasoning indicates it would be hard to build a launch vehicle powered by energy stored in counterrotating flywheels or in compressed gases. Paul F. Dietz dietz@cs.rochester.edu ------------------------------ Date: Wed, 22 Apr 92 20:46:31 -0400 From: dietz@cs.rochester.edu To: space-tech@cs.cmu.edu Subject: factor of 2 ACtually, he didn't drop a factor of two; I just misread his post. But reaching escape velocity + 11 km/s means the stored energy is 2x higher, so the material strength requirements are also higher. That ring is going to have to be made of superconducting unobtainium. Paul ------------------------------ From: sequent!techbook.com!szabo@uunet.UU.NET (Nick Szabo) Subject: Re: Magnetic sails To: neufeld@helios.physics.utoronto.ca (Christopher Neufeld) Date: Wed, 22 Apr 92 14:29:57 PDT Cc: space-tech@cs.cmu.edu Here is an Andrews & Zubrin design for magsail thermal control from their "Progress in Magnetic Sails", from the July 1990 AIAA/SAE/ASME/ASEE Joint Propulsion Conference. In interstellar space, ambient is 2.7 degrees Kelvin, which allows them to operate NbTi and Nb3Sn superconductors at 1e10 and 2e10 amps/m^2 respectively without refrigeration. In interplanetary space they are targeting YBa2CuO7 in the 70-90 Kelvin range without refrigeration, which at that time had demonstrated 1e10 amps/m^2 in thin film samples. (Any progress since 1990?) For the interstellar probe, they chose a "advanced technology" (read: unobtainium :-) superconductor of 1e11 amps/m^2. Two possible outer surfaces: The first is a polished Kapton strip covered with vapor deposited layers of dielectric materials, arranged in alternative layers of high and low refration materials of one quarter wave thickness. Similar mirrors available today have demonstrated reflectivities in excess of 99.9% over wide portions of the solar spectrum. The second is a two wavelength design, where the outermost ten or more dielectric layers have thickness designed to reflect red and infrared, the innermost blue and ultraviolet. Calculations indicate the design should reflect over 99% of the total solar spectrum. For both preceeding two paragraphs, no references or calculations given, just those statements. The thermal control system is a plastic V shape 1" on a side. The sunward surface of the front leg has the dielectric mirror coatings and the anti-sunward side has a high-emissiviity coating (eg SiO2). The rear leg is aluminized on the sunward side to reflect the infrared emissions on the front leg with 99% efficiency. ============== <--- aluminized superconducting cable wrapped w/teflon -------------- <--- rear leg: (outer surface aluminized) / / / <--- front leg: 400 micron Kapton with multiple / dielectric coatings sunward and high / emissitivity coatings on both side (SiO2) The cable is attached to the backside of the rear leg by a Teflon cable wrap. The superconductor outer surface is aluminized to reflect IR while the teflon wrap is an excellent IR emitter. Assuming an emissitivity of 0.8 an an absorbtivity of 0.01 (99% reflectivity) the front leg has an energy input of 0.01 * 1,340 W/m^2 = 13.4 W/m^2 at 1 AU. Radiating 6.7 W/m^2 from front and rear surfaces with an emissivity of 0.8 requires an equilibrium of 110 Kelvin. This means the superconducting cable sees a 45-degree arc of shield at 110 Kelvin radiating 6.7 W/m^2 and a 315 degree arc of deep space at 2.7 Kelvin. The aluminized cable w/teflon cover has absorb/emit of 0.074. Therefore the equilibrium temperature would be 50 Kelvin or less. ------------------------------ Date: Thu, 23 Apr 92 00:03:20 EDT From: John Roberts Disclaimer: Opinions expressed are those of the sender and do not reflect NIST policy or agreement. To: space-tech@cs.cmu.edu Subject: Re: Superconductors > Larry Niven and a few more reputable sources are wrong, in fact almost >backwards. This is probably an attempt to fit the semi-classical > relation between electrical and thermal conductivity, >whcih is more or less a fixed ratio for many fixed metals, in the >Sommerfeld model of a metal (a box of non-interacting fermions). This is >not a close approximation to a superconductor. Argh! - Niven strikes again! I get spoiled reading Heinlein and Asimov, then fall into the trap of thinking Niven put as much effort into getting the details right as they did. His Ringworld is unstable, the attitude rockets probably wouldn't work, the variable-sword wouldn't work as advertised, the disintegrator probably wouldn't work, the galactic core wouldn't be nearly as bright as he described, Beowulf wouldn't have survived even the approach to the 1-mile perigee neutron star trajectory, and so on. If I interpret the entry in the Encyclopaedia Britannica correctly, thermal properties are similar to, but not quite the same as those of the materials at slightly above Tc. It is also of note that there are two main types of superconductors: type II superconductors don't expel magnetic fields in quite the same way as type I superconductors. n Now that I think of it, for the "more reputable sources", I was confusing my phenomena - what I had in mind was superfluids, notably helium 2 (II?) - liquid helium-4 (it doesn't work for helium-3) cooled to below Tlambda = 2.2K. The thermal conductivity of helium 2 is about three *million* times as great as that of liquid helium just above Tlambda. Years ago, I invented a power distribution system that used superconducting cables, surrounded by helium 2 to maintain the proper temperature. With good insulation, I expect the refrigeration units for the helium 2 could be placed many miles apart. John Roberts roberts@cmr.ncsl.nist.gov ------------------------------ End of Space-tech Digest #117 *******************