Saturday, 30 January 2016
Thursday, 3 December 2015
The Neumann Drive has several advantages over legacy solar-electric drive technologies.
Because it uses solid fuel, it saves on the tankage, pumps, pipes and all the other things you need for dealing with compressed xenon in a space environment – it is perfectly possible to just leave the metal fuel rods in a vacuum.
Because it throws the high-energy particles away from the working parts of the drive, it is more likely to have a longer operating life than a given design for Hall Effect or Gridded Ion Thrusters, which push ions either through their grid or past the accelerators, wearing them down over time.
With the right fuels, it is far more fuel efficient than competing models, with specific impulse over 14,000s in the case of magnesium, and around 5,000s with the case of molybdenum, as compared to about 2,000s for a Hall Effect Thruster or about 3,800s for a Gridded Ion Thruster.
But the biggest advantage for the Neumann Drive over the legacy drives is that it can use fuel that doesn't need to be moved up from Earth.
Aluminium is one of those fuels, because much of the approximately 7,000 tons of space junk is aluminium. A large proportion of that space junk is sitting just above geostationary orbit, in the so-called 'graveyard orbit'.
Let's use the old Satellite Business Systems-6 (SBS-6) as an example. SBS-6 was launched on an Ariane rocket in 1990, and between getting to a stable geostationary orbit, and the years of station keeping and then moving to its current graveyard orbit, it used all of its fuel. This took it down to its current, empty "dry mass" of 1.514 tons, and it's likely to be more than half aluminium by mass.
If somebody moved a space furnace into the graveyard orbit above geostationary, and the owners of SBS-6 were co-operative, then you could recycle this aluminium into fuel rods.
These fuel rods will be markedly inferior to those you could bring from Earth, but being already in space means you don't have to pay the very high price for a chemical rocket to get it up there. Even if the results are less than the best, the price-per-pound could be worthwhile. But just how bad/good could they be?
Let's assume a mostly-aerospace aluminium miscellaneous mixed fuel rod is 20% worse in every category than pure aluminium usually is (testing space-grade aerospace alloys is a high priority for us, but we don’t have any results we can release – so let’s simply assume it's even worse than a material that we regard as pretty damn crap to start with).
In testing, pure aluminium fuel rods got a specific impulse of 2,323 +/- 325 seconds and a thrust-to-power ratio of 10.8 +/ 0.5 microNewtons per watt. This means we're assuming you get about 9 microNewtons per watt of power put in the rod, and it'll have a specific impulse of around 1,800.
That means you need about the same amount of fuel to get somewhere as if you'd used a Hall Effect Thruster system, though the HET will need about one-eighth of the mass of solar panels to keep it moving.
A satellite in geostationary orbit needs 50 m/s of acceleration each year to stay there. Lets assume we have a 3 ton satellite. Therefore, we will need 150 000 Newton-seconds of impulse each year to stay in the right orbit.
Let's assume we have 10 kW of effective power to the Neumann Drive. 10 kW at 9 microNewtons per watt makes 90 milliNewtons of time-averaged thrust, or 5.4 Newton-seconds every minute. This makes 7, 776 Newton-seconds every day, so running the system for 19.5 days per year will keep your 3t in orbit.
At a specific power of around 150 watts per kilo, our hypothetical junk-fuelled Orbit Maintenance Tug will need around 70 kilos of Spectrolab ITJ panels for 10 kW of power.
Assuming the Orbit Maintenance Tug is a total of 500 kilos, then with a specific impulse of 1,800s, 250 kilos of Mixed Space Junk fuel rods has a total momentum change over time of 1,217 meters per second, or around 25 years of station keeping, with the fuel obtained from slow-moving objects about 150 km away from geostationary orbit around Earth.
This means construction of the Space Furnace, and its use to recycle satellites in the graveyard orbit will allow the practically unlimited life extension for all satellites currently in geostationary orbit.
This is why the ability to use fuels already in space is the greatest advantage of Neumann Drives over legacy technologies.
Tuesday, 17 November 2015
In science, learning from your failures can be more important than success.
Tin is a really bad fuel for Neumann Drives, but the way it behaves tells us a lot of useful things.
Before anything else, tin is a soft metal with a low melting point - it will melt at 233 degrees Celcius, which is a little hotter than you want to bake chickens at in a home oven.
A Neumann Drive works by passing an electric current through a solid fuel rod used as the cathode, after triggering the arc with a surface flashover from the trigger pin. Thus there are two main things that can reduce efficiency - you let the current pulse run too long, so that the cathode spots run off the edge of the front face of the cathode and arc sideways to the anode ("side-arcing"), or you use a level and duration of current that melts part of the metal fuel rod. This last one means that rather than producing ions that leave the cathode at tens of kilometres per second, you get pools of molten tin that produce droplets and clouds of neutral tin vapour, neither of which move very fast at all.
The image below shows the result of this for tin.
At low currents applied quickly, tin is merely a terrible Neumann Drive fuel, with specific impulses approaching the glorious heights of 600s - about twice as fuel efficient as a solid chemical-fuelled rocket, or about a third as fuel efficient as a solar-electric xenon-fuelled Hall-Effect Thruster. However, when higher currents are applied for longer, performance goes into the toilet and beyond.
Using sixty joule pulses over 300 microseconds, performance drops to below what can be achieved using household sugar and potassium chlorate, the so called 'sugar rocket' of amateur rocketry.
Now, the point is not to make jokes about tin being worse than fuels that are made in kitchens using coffee grinders - it is about the why performance gets worse, which is that parts of the fuel rod itself get melted.
This is why we are so interested in actively cooling the cathode, as then we may be able to maintain the high performance of good fuels like molybdenum as we crank up the pulse rate, moving to higher power levels and thus get more thrust out of the same number of drives.
Friday, 30 October 2015
Bismuth was a material we had high hopes for. It has been used as propellant in other electric spacedrives, notably in the work done by Massey on Bismuth-fuelled Hall Effect Thrusters at Michigan Technological University.
There are reasons to think Bismuth would have been a good fuel for the Neumann Drive – it's a heavy element, non-radioactive and with a low ionisation energy and a potentially useful electron cross-section.
We were hoping for decent results.
We were horribly, awfully, wrong.
In testing, we hit a Bismuth cathode with 15 000 pulses, which eroded just short of fifteen grams of material. At a pulse rate of four per second, that's about an hour of operation.
Magnesium erodes fifty two grams in a month.
Our best results for Bismuth involved hitting it with 100 microsecond pulses, with about twenty joules per pulse – that's half the time and half the energy of magnesium, a fuel which we regard as needing to be treated gently for optimum performance.
These best results involved specific impulse of about 140, which is substantially worse than that achieved by amateur rocket enthusiasts using hydrogen peroxide as a monopropellant.
At higher energies, for example 40 joule pulses over 200 microseconds, it went much worse. We achieved a specific impulse of about fifty, which means it is competitive with the engineer teeing up the lump of bismuth up to hit in the opposite direction of travel with a golf club.
Friends don't let friends use bismuth in Neumann Drives.
Thursday, 15 October 2015
Molybdenum is your go-faster fuel. It is a heavy refractory metal, with a high melting point. It has been identified in the solar system, but it's likely to be a parts-per-million component of nickel-iron asteroids.
You'll need to bring it from Earth, but this is the fuel that gets people to Mars.
In a Neumann Drive, Moly can be hit hard and hit often – it appears to have a linear efficiency pattern that means we can happily use five pulses per second at 125 Joules per pulse for three hundred microseconds, and we're confident that aggressive cooling of the cathode can increase the number of pulses per second. At this rate, the fuel has a power efficiency of 20 micronewtons per watt, and over a month will produce 32,400 Newton-seconds of momentum exchange and use five hundred and twenty grams of fuel to do that.
What this means in practice is that the power demands for a Neumann Drive running Mo are fairly high – you need 3 kilos of off-the-shelf Spectrolab space-rated Improved Triple Junction solar panels will support one Neumann Drive running moly while near to Earth, and about double that if you plan on operating near Mars. Likewise, using 125 joules per pulse means we need a good sized capacitor bank.
If we use 6 kilos of Spectrolab ITJ panels (to allow for degradation over time and efficient use as far out as Mars) and share power management resources among Neumann Drives so each has a "slice" of three kilos of shared capacitor and two kilos of shared inductor per Neumann Drive, and each Neumann Drive thruster unit masses four kilos by itself, then each Neumann Drive optimised for molybdenum will have a total mass of fifteen kilograms, including shared supporting systems.
Add six kilograms of fuel, which should be enough for just over 12 months of continued operation.
We therefore are committing 21 kilos to create 32,400 newton-seconds of momentum exchange each month for a "burn period" of just over a year.
If we have 100 of these (adding to 2.1 tons) and 700 kilos of structure, cooling, flight computers and payload; then the spacecraft adds to a total of a fully-fuelled "wet" mass of 2.8 tons, and an unfuelled "dry" mass of 2.1 tons. When we are in enough sunlight all the time, the drives create 3,240,000 newton-seconds of momentum change each month, which will represent 1,150 metres per second per month of acceleration while the craft is fully fuelled, and 1,542 metres per second per month of acceleration when running on empty.
For the sake of a first-cut approximation, let’s assume an average acceleration based on average mass of the craft, of 2.45 tons – what does it do running Moly when fully loaded with half the fuel tank gone.
With 3,240,000 newton-seconds pushing 2.45 tons, you get an acceleration of 1.32 kilometres per second per month.
Low Earth Orbit to Mars orbit needs approximately 15 kilometres per second of delta-v when using low-thrust solar-electric propulsion. We have an average acceleration per month of 1.32 km/sec, and 12 months of fuel, we have a delta-vee budget of 15.89 km/sec.
Clearly, this is for a people-to-Mars mission, which introduces some complications. Assuming your people are leaving from Earth and you want them in good condition at the other end, you probably don’t want to be taking them through the Van Allen Radiation Belts via solar-electric propulsion, because even if you avoid the really dangerous parts, spending time in the radiation trapped by Earth's magnetic field isn’t good for humans.
A better plan would be to send the humans directly to a transfer station at L5 via a fast chemical rocket, spending the fuel to keep them healthy, while their stuff goes in advance from where it was dropped off in LEO
If the rocket that got their stuff from the surface drops their it well in advance to some conveniently-located International Space Station or other, then it can be moved slowly by Neumann Drive to a second station at L5. The humans go directly to the second station via a much faster chemical rocket - something that spends the extra fuel to take three days in the Van Allen Belts rather than three months.
As an aside, this second station can and should be the mark one of any Mars Orbital Habitat, going through its multi-year shakedown cruise somewhere it can get spare parts delivered. Once most of the bugs are shaken out, the mark two gets built, assembled in orbit and moved to Mars orbit.
The humans are then reunited with their stuff, and then take the moly-fuelled Neumann Drive craft from the second station to Mars, but from a point where we have pretty much escaped Earth's gravity field. This will take less kilometres per second of delta-vee, which means we can shave the fuel down, which gives us lower mass, which gives us more acceleration and so on. If you assume we need 10 kilometres per second, then you only need 4 kilos of moly per drive, which drops average craft mass from 2.45t to 2.25t, sending average acceleration goes up to 1.44 km/sec per month, which means a touch over 7 months from the transfer point to Mars orbit.
If you're happy to spend the ton or so of chemical fuel to get each human from Earth to the transport point fast, then the humans will experience an eight month trip, split between chemical propulsion to get them through the Van Allen belts and solar electric propulsion to get them to Mars orbit.
Recent work on the magnetic exhaust nozzle has improved this situation, as less material is being used per pulse to create the same thrust, but the higher atomic weight of moly seems to make the magnetic exhaust nozzle less effective than other fuels we've tried.
Aggressive cooling of the cathode should be possible, and let us jack up the number of pulses per second when we have the surplus power. You need roughly double the mass of solar panels to create the same number of watts near Mars, so if you match your solar requirements to what you'll need near Mars, then you have a bunch of spare power closer to the Sun. But we aren’t even at benchtop testing for that. Theoretically, it should mean the above design gets to use 10 pulses per second near Earth, and 5 near Mars, which means you should use the same amount of fuel but get there faster. How much faster, we don’t know, as we haven’t melted enough cathodes to find out, but we're working on that.
Regarding power systems, the Spectrolab ITJs are actually their entry-level models, and you can get better from them, which will shave the base solar need from six kilos to five. For people-to-Mars, its probably worth buying their expensive ones (if you need to ask the price, you cant afford it). Azurspace are achieving 350 watts per kilo specific power with their cells, but that's cells not panels, so we did the numbers using Spectrolabs' panels. Interesting work is also being done by various parties with thin-film solar on low mass substrates, and there’s mutterings about a kilowatt per kilo being possible soon.
We're also assuming standard capacitors. Interesting things are being done by Skeleton with ultracapacitors, and we're confident graphene-based ultracaps will continue to improve.
Tuesday, 6 October 2015
Magnesium is an interesting fuel for several reasons. The first one is that it is a metal that, because of it's combination of light weight, high strength and ability to be easily formed, is extensively used in aerospace applications. The second is that it has been identified in various places in the solar system, usually as the mineral Olivine. The third is because its specific impulse – how much momentum change you can get out of a given mass of fuel – is utterly insane. All of which combine to make Magnesium one of the best fuels around for the Neumann Drive.
This is the fuel that gets you to Mars and back on a tank of fuel, and we will now give you the numbers to show you how.
In a Neumann Drive, Magnesium needs to be treated fairly delicately – it appears to have a “sweet spot” of four pulses per second at 40 joules per pulse. At this rate, the fuel has a power efficiency of 11 micronewtons per watt, and over a month will produce 7,200 newton-seconds of momentum exchange and use just 52 grams of fuel to do that.
What this means in practice is that the power demands for a Neumann Drive running magnesium are quite low. Three kilos of off-the-shelf Spectrolab space-rated Improved Triple Junction solar panels will support one Neumann Drive running magnesium... even in Mars orbit. Likewise, only needing 40 joules per pulse means we can get away with a smaller capacitor bank than when using some other fuels.
If we use four kilos of Spectrolab ITJ panels to allow for degradation over time and efficient use as far out as Mars, and share power management resources among Neumann Drives so each has a "slice" of two kilos of shared capacitor-bank and two kilos of shared inductor per Neumann Drive, and each Neumann Drive thruster unit masses four kilos by itself, then each Neumann Drive will have a total mass of twelve kilograms, including shared supporting systems.
Add six kilograms of fuel - which is enough for 115 months of continued operation, or just short of two five-year missions.
We therefore are committing 18 kilos to create 7,200 newton-seconds of momentum exchange each month for a "burn period" of just short of ten years.
If we have 100 of these, this adds to 1.8 tons. Then add 700 kilos of structure, cooling, flight computers and payload, then the final spacecraft has a fully-fuelled/"wet" mass of 2.5 tons, and a empty/"dry" mass of 1.9 tons.
When we are in enough sunlight all the time, the drives create 720,000 newton-seconds of momentum change each month, which will represent 288 meters per second per month of acceleration while the craft is fully fuelled.
Low Earth Orbit to Mars Orbit needs approximately 15 kilometres per second, when using low-thrust solar-electric propulsion. At a minimum of 288 meters per second per month of acceleration when fully loaded and 115 months of fuel supply, we have 33,120 meters per second of delta-vee budget (in reality, it will be better than this, as the craft will accelerate better as more fuel is used and it gets lighter. To keep the maths simple, assume we picked up around 300 kilos of moon rock at Phobos or something).
If we are using no fancy navigation (like the Mars gravity assist that the Dawn probe used to help get to Ceres, or the Earth swing-by that Hayabusa used on it's way to 25143 Itokawa), this shows a return trip to Mars orbit, Deimos or Phobos is easily achievable with these assumptions about engineering.
Recent development of the magnetic exhaust nozzle has improved this situation, as less material is being used per pulse to create the same thrust.
Tuesday, 29 September 2015
Titanium is an interesting fuel for a couple of reasons. Firstly, it's a pretty popular aerospace metal, so there is a fair bit of it floating up in Earth orbit ready for recycling. Secondly, as far as the base Neumann Drive goes, it is a well-behaved light refractory metal that does what it should. Thirdly, it responded really, really well to the magnetic nozzle, achieving marked increases in efficiency.
During the testing process in 2014, we hit a Titanium cathode with 460,000 eroding pulses, which sent approximately 3.8g into the exhaust. As far as fuel efficiency goes, Titanium had the best results using 30 joule pulses over 100 microseconds at 4 pulses per second, which resulted in about 2.5 Newtons over that time, with a specific impulse of about 4200.
Once we got the magnetic nozzle attached, the results changed for the better – but to compare them, we need to move to the work we did on 'clean', un-eroded cathodes. A brand new, “clean” cathode temporarily gets better results, but we generally don't like talking about those results because nothing stays new in space for long, and a few hours of solid use will change a previously clean cathode to a used one.
A clean titanium cathode running 200 microsecond pulses at 50J per pulse will produce 3.75N of thrust over the length of the pulse, at around 3000s of specific impulse, without a magnetic nozzle. Adding the nozzle to the system improved the performance under similar conditions to producing about the same amount of thrust, but at around 4500s of specific impulse. This represents a 50% improvement in efficiency, with no added power or fuel required and with no reduction in thrust.