I am currently on a plane on my way to Shenzhen by way of San Francisco and I was going to write something about business and science but who has time for that. Instead lets talk about asteroid mining. While an extremely syfy concept, my interest in asteroid mining is one of practicality. All deep space missions up till now have been done in the name of pure scientific discovery. While for me thats reason enough, in the world and funding environment we live in it is not. With the recent government shutdown, almost all of NASA was deemed nonessential. So if we as scientists not just rocket scientists want to continue pushing the boundaries of science at any acceptable pace, we must take control of funding
What exactly are we mining here?
Metals. The relatively small 2.3km wide asteroid (6178) 1986 DA contains an estimated 10,000 tons of gold, 100,000 tons of platinum, 10 billion tons of iron and a billion tons of nickel. And according to Wolfram|Alpha thats $384 Billion of Gold, $4.2 Trillion of Platinum, $269 Billion of Iron and $20 Trillion of Nickel. For a grand total of $25 trillion US dollars. Now the economists in the bunch are going to be saying “Surely if you drop 20 times the amount of platinum produced ever or 5x the amount of gold produced annually onto the market you’ll make the metals worthless” To which I’ll respond, well obviously you wouldn’t sell of your entire stock in one year. Plus demand for gold and platinum isn’t going anywhere but up, especially since we’ll likely run out of both within the century (and unlike mined diamonds, rare metals have an ever growing demand courtesy of the electronics industry). Gold is much more valuable as a component than locked away in a bank.
Now on to the fun bits.
In this post, I am only going to be talking about the first phase of Asteroid Mining, which luckily for us also happens to be the first phase in terrestrial mining; and that is Surveying. Of course, now is probably a good time to tell you that I am not a geophysicist or really any kind of physicists [in fact I’d be very surprised if I don’t receive at least a few emails pointing out what I got wrong] and everything I saw should be taken with a grain of salt.
On earth surveying starts with satellite images and based on those images teams do ground surveys. NASA takes a similar approach to how they study mars, send orbiters to do flybys, and then rovers to take samples. If we had the funds of a government or a large corporation following a similar approach would also work for surveying potential asteroids. But it’s much more fun to innovate around a problem vs throwing money at it. So lets pretend (is it pretending if its real) like we don’t have billions of dollars or even millions, and figure out the cheapest system that would still get us the quality of surveying data we need to move on to the next phase of asteroid mining.
We know we need a satellite, and that we’ll have to launch it into orbit around our target, which means we’ll also need thrusters and fuel on the satellite to move it into orbit.
The Satellite Components
There are some pretty cool experiments around using Android phones and other off the shelf components for use within satellites. However they have currently only been tested within the protection of Earths magnetic field and were short missions. A satellite traveling into deep space for weeks possibly months would face much larger doses of solar radiation as well as run into cosmic rays, have to survive the Van Allens belts plus the odd solar flare. So bootstrapping a satellite together from radio shack isn’t going to work, It is going to have to be custom built from scratch.
The primary sensor we’d want is a hyper spectral imaging camera. Which at its most basic is an objective lens that directs light through a hyper spectral filter onto a CMOS sensor. Most hyperspectral cameras are very expensive and large because they use discrete objectives and prisms. For example a relatively small model by Middleton Research is 2.7kg which might not seem like a lot but when you’re launching things into space every kg costs $5k (cost to reach LEO on SpaceX, going past earth orbit will be more expensive). Luckily for us there have been recent advances into integrating the filters onto the sensor itself. The company leading the way on this front is IMEC (see their prototype above).
For the sake of this post lets base our computing requirements around the CMOS sensor used in IMECs imager above, the CMV4000. Which outputs each frame at 5MB and has the capability of outputting 180fps. Currently the best option for memory is Everspin Technologies MRAM modules which recently developed 64Mb modules. For storage of the data, we would use Flash since you can get radiation hardened modules up to 64GB. There are a variety of rad hardened processors to choose from Atmel, PowerPC, since we’re not doing any photo processing in space their relatively low computational ability (133 and 25Mhz respectively) isn’t a problem.
To broadcast data and control the satellite from a few million miles away, we’ll need a combination of low-gain, high-gain and UHF antennas. The omni directional low gain will allow a certain level of control and data to be sent and received at all times but only in the low b/s rate. The high gain in combination with a dish will be the primary way image data is sent back capable of Kb/s . Bigger is better on this front, except bigger means larger payload to launch, so it might be worthwhile to experiment with ways to get a dish that could be unfolded after launch. And as a backup it’d be worth it to launch a secondary satellite within UHF range of the imaging satellite, to ensure fidelity of the signal.
To estimate the potential speeds, I looked at what the Mars Science Laboratory (aka Curiousity) is capable of. And with a 15Watt radio, it’s able to get 15b/s with it’s low-gain antenna and 32kb/s with it’s high gain directional antenna. Now looking at the Mars Reconnaissance Orbiter which uses a 100Watt X-Band amplifier and a 3meter dish, we see it’s possible to get speeds up to 6 Mb/s.
The Deep Space Network, the collection of satellites and dishes that relay data to and from space is really quite interesting and I need to do a lot more research into satellite broadcasting and antenna design. If someone wants to donate a copy of the Space Antenna Handbook to the Brightwork library.
Everything is relatively low power, the CMOS only uses 600mW, the computer board 8W at peak and the largest power consumer is going to be the power of the radio which even if we transmit at 100watts is not going to bring our power usage above 500 watts. Using the latest Triple-junction cells we can get 500 watts(based on the solar energies at mars distance) with 35 square feet or 2 panels of 6′ by 3′. In reality, we would probably be okay using a lower powered radio to drop energy usage to 100watts (which would only require 7ft2 of solar panels). And more importantly a lower power draw would allow for a smaller battery pack since Lithium ion batteries weigh 1kg per 265Wh.
The thrusters serve two purposes, orbit insertion and attitude adjustment. Hydrazine thrusters seem to be the go to as thats whats on the Voyagers as well as the more recent Mars Reconnaissance Orbiter. The downside is that they require large amounts of fuel which adds to the weight. The lighter alternative would be Ion thrusters like those used on SMART-1 but they have the downside of requiring kilo-watts of power and not being able to accelerate as rapidly as chemical engines.