satnet

Members
  • Content Count

    101
  • Joined

  • Last visited

Everything posted by satnet

  1. Since each sub-pixel has its own emitters you wouldn't require an analog connection. It would be a digitally addressed grid just like an LCD and could natively use the same signals it could. Given that almost all of the advantages of this technology are shared by OLED (high contrast, no backlight, low response time, wide viewing angles) I doubt we'll ever see this come back. There just isn't enough advantage to displace the now reasonably established place in the market for OLED. The only advantage I even see is that the phosphors will probably last a bit longer. There wouldn't be an advantage to going back to analog signals. GPUs have always created a digital image and only turned it into analog signals if that is how it was to be output. The main work of a GPU is all of the parallel processing needed to build an image, in effect all of the light interactions that occur in virtual 3D space with multiple virtual lighting sources and virtual objects. The actual creation of display signals after that 3D processing is turned into a 2D image is a small part of a GPUs job that has had the option of both digital and analog output from the same GPU for several decades.
  2. Musk is correct that is the point of a nozzle (though that is immensely simplified, and not a new concept since it is rocket design 101). Basically a nozzle takes high temperature fluid and turns it into high speed fluid with large amounts of kinetic energy (see de Laval Nozzle). Temperature is random particle motion in every direction, while kinetic energy has a common velocity vector. In other words a nozzle is what forces the randomly moving particles to move (mostly) in the direction we want them to. A simple cylinder isn't going to do much in terms of shaping the direction of a fluid's velocity vectors, though you can certainly create magnetic fields with the convergent-divergent behavior you need, just with more complex shapes. As far as high thrust is concerned, the answer is that you can have a higher thrust plasma thruster such as VASIMR, though even at max thrust it is only 5 Newtons. This is much higher than the 0.327 N of the NEXT ion thruster (itself a noticeable improvement over prior ion engines), but a long way from the 110 kN of an RL-10 hydrolox engine and a really long way from anything that will launch you from the sea-level on Earth. The high energy requirements of a high thrust plasma/ion engine are difficult to meet. You also have the usual problem of needing a lot of reaction mass for a high thrust engine which increases the mass you need to take with you. Plasma/ion engines are really good at accelerating exhausts to high velocity which makes them poorly optimized for high thrust operation. It is more energy efficient to get more thrust by throwing more reaction mass than by increasing velocity, but increasing the reaction mass means you have more mass to move so rockets are actually better off devoting energy to higher velocities. Plasma engines are doubly impacted when trying to increase thrust since you need to increase both reaction mass and energy generation/storage mass (chemical engines have the advantage that their reaction mass is also their energy storage mass). Antimatter gives you really dense energy storage and a lot of energy to work with which does help a lot, though you still need to deal with waste heat and you still need reaction mass. You could do a higher thrust plasma engine using antimatter as an energy source, but the rocket equation still holds true and you would have an exponentially smaller spacecraft if you take the high exhaust velocity, low reaction mass, low thrust approach.
  3. Another negative impact of the square-cube law is that expander cycle rockets are limited to ~300 kN of thrust. This is because the surface area over which you can extract heat grows slower than the volume you need to fill with fuel and at some point you can't extract enough to run the pumps to fill the volume.
  4. Floating point is already logarithmic, and already represented by a pair of integers (plus a sign bit). The most common representation is IEEE 754 which defines several formats, but consists of a sign bit, significand, and an exponent with a pre-defined base. When you use a float data type it is (usually) a 32 bit representation with 1 sign bit, 23 significand bits, and 8 exponent bits with a base of 2 (making the exponent the whole number portion of log2(N) ). Converting this into a base 2 number looks like this N = (-1)(sign bit) x significand x 2(exponent) which is very similar to scientific notation except in base 2 since binary computers naturally operate with a base of 2. IEEE 754 does define a base 10 representations as well, though they are recent and if anyone uses them I'm not aware of it. Basically you won't save any time because this is already what we do. On the other hand you have managed to come up with a perfectly viable approach with a long history of working. Base e is not likely to work well. Since it is an irrational number a computer needs to approximate it to some number of digits. Computing the log base 2 of a number can be computed with a fairly simple circuit, log base 10 is more complicated yet reasonable, and log base e is just an awful mess. P.S. I'm glossing over some details. There is more to this than is relevant to the discussion.
  5. The SpaceX environmental impact study includes a third party simulation of the exhaust products (page 169). Per the summary: Calculations were performed to estimate the far-field exhaust constituents of the SpaceX Raptorliquid oxygen-liquid methane (LOX-LCH4) booster rocket engine firing under sea-level conditions. Although the exit-plane exhaust is fuel-rich and contains high concentrations of carbon monoxide (CO), subsequent entrainment of ambient air results in nearly complete conversion of the CO into carbon dioxide (CO2). A small amount of thermal nitrous oxides (NOx) is formed, all as NO. The CO and NO emissions are predicted to be less than 0.024 lbm/s each, per engine under nominal power (100%) operation. No soot is predicted to be generated by this engine cycle. The CO and NO emission rates for the Super Heavy has been estimated to be no more 0.788 lbm/s each. The predicted sea-level CO and NO emission rate for the Starship upper stage are estimated to be no more than 0.168 lbm/s each. Table 3: Thrust Chamber Nozzle Exit Species Mass Fraction from VIPER Simulation Species Mass Fraction CO2 0.39950 H2O 0.41333 CO 0.12071 O2 0.054752 H2 0.007462 OH 0.0035882
  6. I know this isn't the Raptor, but Scott Manley did a detailed video on the F-1 engine startup procedure that was pretty interesting.
  7. The EM Drive has been reasonably well debunked at this point. Dresden University of Technology did the tests that debunked it for most of the mainstream. They used a rig with very accurate force measurements and the ability to reorient it. Based on their observations most have concluded that the thrust measured is just inadequately shielded EM fields interacting with Earth's magnetic field. There are a number of tests that back this up, but probably the most convincing is that when they attenuated the radio waves going into the resonance chamber the thrust did not change. A far more interesting question is whether it was worthwhile investigating a fringe hypothesis like this with public funds. I tend to say yes since there was enough evidence that something was going on to warrant looking and we can't be closed off to all things contradicting established theories, though I don't think continuing investigation is warranted (at least not with public funds). Having said that the EM Drive always fell in the "extraordinary claims require extraordinary proof" category and it never delivered more than vague indications until we arrived at a more plausible explanation.
  8. Putting fuel into a rocket engine does not create momentum. Momentum is conserved, you can't create it though you can exchange it in interesting and useful ways. The overall system has the same momentum it started with. Basically if you were to take the vector sum of the momentum of the rocket and the rocket exhaust it would add up to its starting momentum (at least in a vacuum where we can treat this as the total system we're analyzing). A rocket doesn't work by creating momentum, but by dividing it in a controlled manner between the remaining mass of the rocket and the exhausted mass so that the velocity vector of the remaining rocket mass increases in a specific direction. To recapture all of the propellant you must undo this which means that the rocket and exhaust must end up with a net velocity vector of zero and as long as they remain together this means the rocket will not move. You could of course recapture a portion of the propellant and exhaust some of it. You would decrease your change in velocity relative to just exhausting all of it, but it would give you a variable thrust engine with consistent cooling characteristics. This gives us the only reason I can think of to pursue this dubious enterprise. You could create an engine that had a relatively high initial thrust until it ran low on propellant which then switched to recapturing its working fluid (no longer propellant, now coolant) and used the waste heat in a photon rocket to continue accelerating for as long as you have nuclear fuel (at an obscenely low rate of acceleration).
  9. Chemical rocket engines have very high thermal efficiency, bordering on ideal. An efficient chemical rocket has about 70% thermal efficiency. Combined gas and steam turbines used in some maritime applications are about 65% efficient. Jet engines have about 40% thermal efficiency. Gasoline engines are about 30-35% thermally efficient, diesel can reach about 40%. The low efficiency of piston engines is the reason electric cars make sense. Batteries have about 1/10th the energy density of hydrocarbons, but are much more efficient at converting that energy into usable work (90% vs 35%), so you can still get a useful range out of them in spite of the huge initial disparity (though it is large enough that piston engines still have longer ranges unless you devote more mass to batteries). Where rocket engines do poorly is overall propulsive efficiency, which includes converting that usable energy into propulsion. This is because rockets need to carry all of their reaction mass with them. The most efficient way to propel something is to move a large reaction mass with a small, but opposite velocity. This is because momentum is m*v meaning an increase in either mass or velocity has an equal change to momentum, but kinetic energy is 1/2*m*v2 meaning that increasing velocity requires significantly more energy, so the most energy efficient route is to use larger reaction mass and smaller velocity. Rockets can't use a large reaction mass for higher propulsive efficiency because this increases the amount of mass that needs to be moved which negates the advantage of using a large reaction mass, so it is ultimately more efficient for a rocket to devote that usable energy into a high exhaust velocity and reduce the amount of exhaust mass. Aircraft, boats, and cars all have the advantage of a constant supply of reaction mass they don't need to carry so they can devote their usable energy to moving a large reaction mass. NOTE: All of these percentages are approximate and most are on the high end of the range.
  10. LightSail 2 has successfully demonstrated solar sailing. It has raised its apogee by 2 kilometers using a solar sail for thrust, a momentum wheel for orientation, and electromagnetic torque rod for desaturation. http://www.planetary.org/blogs/jason-davis/lightsail-2-successful-flight-by-light.html
  11. Actually there is an interesting proposal to use in space construction to build something on this scale. I'm not sure how likely it is to actually be built, but it does allow for some interesting possibilities.
  12. In this case voltage, current, and resistance vary significantly with time and you also have to worry about impedance since we're dealing with an AC circuit with reactive components (see https://en.wikipedia.org/wiki/Ohm%27s_law#Reactive_circuits_with_time-varying_signals). I'm pretty sure the voltage will jump as high as it is going to get when you disconnect that switch. While it is superconducting you won't lose energy to resistance (at least through the inductor), but that will change when you quench (lots of energy dissipated as heat). That's assuming your capacitor doesn't burn out the moment you disconnect your supply. You can get the conditions to power fusion using inductors and capacitors to boost voltage since Farnsworth-Hirsch fusors commonly use high voltage flyback transformers (with ordinary inductors) to drive them. Unfortunately high voltage fusors generally don't reach an energy positive point. They are useful as neutron sources and research, but not useful for doing net positive work. The electric and magnetic fields of the plasma would change very quickly with time. This is one of the major hurdles of fusion, once you get it started the plasma generates its own electromagnetic fields that counter the beautifully orchestrated EM fields you used to get them to fuse in the first place.
  13. Essentially you're talking about a boost converter, using the inductor as both the energy storage and the switching element. Using magnetic quench instead of a transistor for the switch is interesting, but probably not terribly efficient. Generally a boost converter is switched at a very high rate to minimize losses, which wouldn't be practical with magnetic quench since the time to re-cool the superconductor would be very long.
  14. This type of station probably wouldn't be for satellite refueling. It would be for launcher fueling. For example you launch into a rendezvous orbit with minimal fuel left, dock, refuel and then continue to GEO, Lunar, or Mars orbit. This exponentially decreases the size of your rocket (or exponentially increases your payload capacity). If you did want satellite refueling, NASA has researched refueling satellites before and they did operate on the assumption that they wouldn't be provisioned for refueling which was proven to work on an ISS mission (Robotic Refueling Mission), though it wasn't a real satellite just a representative example. Chances are you wouldn't have the satellite come to you, but instead send out a tug with just as much fuel as you need to resupply it (see Space Infrastructure Servicing). Methane will be used by Vulcan's first stage (ULA), New Glenn's first stage (Blue Origin), and of course Starship (SpaceX). However, you do have a point since only Starship's orbital stage is methane so only it could benefit directly from methane in orbit. Having said that generating hydrogen from methane is relatively easy (pretty much all commercial hydrogen is produced this way anyway). You would need to liquefy it, which is difficult, but if you have the lead time, power, and cooling systems entirely doable. ULA's ACES upper stage is planned to have refueling provisions which would be able to benefit from a hydrolox refueling station. A fuel depot can afford the extra mass for active cooling, because they don't need to worry about their dry mass/wet mass ratio to the extent a spacecraft maximizing delta-v needs to. The real problem a fuel depot needs to overcome is economic. They need to have enough clients to justify the large up front cost and ongoing overhead. This means either it needs to be cheap to launch and operate or you need a high volume. If we go to the moon and stay for a prolonged period (or even just have an active tourist industry) you probably have the volume to justify it, but until then there probably isn't enough GEO, Lunar, Mars orbit launch volume to make this economically viable (ignoring the case where it is heavily subsidized, which it probably could be).
  15. Definitely a dubious prospect. The engineering challenges are a huge hurdle and even after you solve them, the things it could launch force them into a pretty small niche. It does have one advantage over a lot of other non-rocket space launch systems: you can conceivably change your launch azimuth. A gun or linear accelerator approach would almost certainly be a fixed installation, but this one could change the release point or even rotate the whole launch system into the proper orientation (since it must be movable). Still like most non-rocket space launch systems it probably makes a lot more sense on an airless moon where you don't have an atmosphere in the way and you can scale down the system to something you can practically build because the orbital velocities are lower. Scott Manley did a video on them about a year ago. No huge surprise that a lot of the potential problems mentioned earlier in the thread were brought up, but he added few more and threw in a little bit of interesting history relevant to the topic.
  16. That code makes more sense, but you're still only calculating heat transfer for one neighbor. You have a constant heat source at x=0, which you have effectively modeled by starting at x=1 in your loop (once you have an x-1 neighbor calculation the x=1 cells will draw from the x=0 cells, but the x=0 cells will never change their value making them a constant source). You are effectively giving y=0 a constant temperature by starting at y=1, but y=0 is still participating in heat flow so it is not an insulator. I would assume you would calculate from y=0 to y=99 and use boundary checks to just not calculate heat flow in the out of bounds direction to simulate an insulator rather than have a row of constant temperature cells. When printing a double you should use %f or %e. For this I would use %e (scientific notation format) probably with a precision specifier (i.e. %.3e, which will give you a number with 3 numbers after the decimal like 1.234e+1). You might find %f more readable in early stages, though since it generally won't use an exponent. The %d specifier is for integers.
  17. The Achilles heel of most of these air breathing engines (ram/scram jets) has been that their operating range it too narrow to go from 0 to orbital or even a significant fraction (by everyday flight standards they cover a huge range, but space flight requires 0 to 7.7 km/s which is a great deal larger). Within their operating range they have a higher overall propulsive efficiency, but to cover the whole range you need multiple engine types which means more mass for engines and less for payload often to the point of being sub-orbital. There are proposed dual-mode scram jets that can work at lower speeds, though I don't think they've been designed or demonstrated and I think they just get into the ram jet territory not down to subsonic ranges where turbofans dominate. Nuclear rockets tend to have a low thrust to weight ratio which usually means they are limited in the design to space or late stage sub-orbital where you need delta-V more than thrust. Of course this is also a problem for scram jets so you now have to low thrust to weight ratio engines which is a real problem. You might be able to use the nuclear reactor as the heat source for a dual-mode scram jet, which would at least help keep the weight down assuming the heat transfer mechanism isn't as heavy as the fuel injection system (and it could very well be as heavy or heavier negating the benefit). Of course as long as we're talking fission and not fusion, as @ARS pointed out you have the environmental and shielding issues to deal with. Isaac Arthur just did an episode on space planes and talks about ram, scram jets and a few other things. He has other episodes on nuclear rockets, though it has been long enough I don't recall if any would be relevant.
  18. Unless your instructor gave you an existing library it is up to you to create the model. I don't know quite enough thermodynamics to know all the properties you'll need to track (temperature and density are probably a given, but I doubt that is all), but in general you'll need something to store the properties of each cell, in this case probably a C struct. If you are using a simple grid a 2d array of these structs would probably make sense where each index into the array represents your spatial step unit in either the x or y dimension. This is a simple model, but it sounds like your professor gave you a problem where a simple model is sufficient. Then as @Cunjo Carl indicated for every temporal step you update the cells in your array representing your new model using you old model array, display it, swap it into your old model array, then start again.
  19. You also have a huge change in center of mass while trying to maintain the same center of thrust. Traditional rockets only shift it vertically so the thrust vector still passes through the center of mass. You could fire the engines on both craft, but that means the X-33 has less fuel post separation. All of the problems are probably solvable, but you need to answer the question: in the face of these challenges how is this better than BFR or something like it?
  20. Actually Arduino uses C/C++, but has a pre-processor that auto-generates some of the things you normally need to do by hand (i.e. generating header files). It will gladly accept ordinary C/C++ in addition to the simplified version that only works with this pre-processor. There are a couple of projects that use Arduino's in quadcopters, which means that there is code out there for interfacing with sensors and making real-time adjustments to maintain stable flight. Unfortunately I checked my old links and they seem to be defunct. Having said that that I'm sure there are a few projects covered on Hackaday that would help you out. One thing I'll point out is before you do this is check all relevant laws (nation, state/province, city). Depending on where you live, building a fully guided rocket may put you on the wrong end of some serious arms control or flight regulations. Arduino is a pretty good choice. It was designed for people who want to do things with electronics, but aren't necessarily wanting to become software or electrical engineers, so it tries to make things as simple as possible. If you really want to use python (or any other language that isn't tied to a specific platform) you might want to look into a Rasberry Pi or a BeagleBone. They are both small boards designed to interface with hardware and run linux (which means full support for python, c, c++, or just about any other language) though they tend to lean towards one language or another. I haven't used a Rasberry PI, but I believe it has good python support. I have an old beaglebone I haven't touched in a few years. Oddly they use javascript for their preferred language, but if I recall correctly you can interface with most things via the linux filesystem, so any language should work, though performance may suffer.
  21. You could probably design a system that was safe. For fusion the exhaust isn't really a problem being mostly helium, though some reactions do produce radioactive atoms like tritium. You probably don't want a fusion torch drive because of the electromagnetic radiation (particularly UV and X-rays), but one where the reaction is contained should be something that can be safe (assuming you can afford the weight of a reasonable amount of shielding). All of this assumes of course that net-positive fusion can be miniaturized once we work out how to achieve it. Tokamak's that might achieve net-positive fusion are building sized (ITER), so that design is probably out. There are reactor designs that are smaller like the Lockheed Martin Compact Fusion Reactor, though I'm a little dubious being such a far departure from existing efforts. You do have the problem of neutron activation of the reactor material itself, though most proponents consider that manageable with the right materials. Assuming you aren't using tritium or another naturally radioactive fuel you don't have the problem of raining radioactive fallout in the case of a RUD event. The challenges for a fusion based in atmosphere rocket are going to be public perception and getting the size/weight down to the point where it is practical. Even nuclear fission can avoid a radioactive exhaust with the right design. The problem there is you if the rocket explodes you now have an uncontrolled nuclear material falling from the sky. You also can't really turn it off (just slow it down) which means you need to be cooling it all the time. This is another advantage of fusion since it can be shut down when not in use.
  22. They might use a mission designation instead of the ship name (i.e. CRS-42 instead of Heart of Gold) during a mayday. On the other hand I skimmed some of the Apollo 10 transcripts and they seemed to handle ships named Snoopy and Charlie Brown without too much drama, and they did use them as handles during radio transmissions (along with 10, Apollo 10, and Snoop). This did lead to a rule change at NASA, but apparently Apollo 16 called their command module Casper anyway. It isn't anything new, though SpaceX's situation does give them the freedom to use as much whimsy as they like (or at least more than most).
  23. As part of the tests they ran on snoopy the docking port was ejected, so it wouldn't be an option.
  24. An SSTO would have an altitude compensating nozzle of some kind. You are basically trying to match pressures between the exhaust and the ambient atmosphere, which is changing as you ascend. As the outside pressure drops the ideal nozzle size grows until it is infinite in a pure vacuum (in the real world we must settle for something less than infinite, fortunately the point of diminishing returns fits within a interstage). A mechanically adjusting "jet" style nozzle which allowed it to expand while in a vacuum would be an option, though generally for the reasons @Dundral mk2 mentioned the approach is usually a different geometry rocket that doesn't require significant mechanical adjustment. If you could afford the weight, had materials that could withstand the stress, and your SSTO spent a non-trivial time in atmosphere it might make some sense to have a mechanically adjustable nozzle (though probably not). It would probably be much longer and expand much wider than anything you would see on a jet, but would visually resemble it. Basically with a little hand-waving you could have a nozzle like that, but the chances of that ever being what we see in the real world are slim. Scott Manley's video on nozzles is relevant (and as usual excellent):
  25. They could land it on a barge in the Atlantic and ship it back. You still need to move it by sea, but you can avoid the canal, though at an average of US$54,000 for an entire ship passing through the canal they might just add that to the launch cost and still be ahead of their competition by a fair margin (assuming it delivers an appreciable reduction in launch costs). They could also land it on the west coast of Africa and then ship it back (which might be cheaper than going through Panama, but obviously not faster). You could also land on west Africa, refuel, then do a sub-orbital hop back to Florida, which would be very expensive, but extremely fast and a good demonstration of rapid reuse. Since Starship will be returning with at least some dV they can boost back towards land after going out to sea. If they start re-entry over land, but at a point where even a post-breakup ballistic trajectory would take the debris out to sea they might convince NASA and the FAA to allow it, which would minimize the boost back dV. If they aggressively use aerodynamic drag and whatever fuel they can afford to slow down they might be able to maintain a trajectory that ensured the Starship would land in the ocean without going too far away from the landing zone then fly back towards the landing zone and trajectory similar to a Falcon 9. My best guess is that they'll go with the US west coast landing and trip through the canal until either launch cadence or competition forces them to look into one of the other options.