wumpus

They may have to, but they also might simply throw away a "used up" ITS after n flights. Perhaps some modules will have longer lifecycles than others and they will detach and deorbit certain modules and replace them, leaving other older parts to go back to Mars. Certainly, the lion's share of the cost will getting the thing from Earth to LEO, so they don't want to do that any more times than necessary (regardless of how many times the BFR can be reused). Landing a ITS on Earth seems like the option they will avoid as long as they can.

How much of a satellite's cost is due to weight minimization? If you could simply stroll over to the Digikey website and build your satellite out of off-the-shelf parts (plenty of parts from newegg and microcenter) and simply weld and enormous heatsink/radiator to it, your costs are likely to go down. You would also presumably need multiple birds in flight because your reliability would be nothing like the billion dollar birds in use. This almost certainly isn't an option for GSO satellites: there are only 120 slots or so around the equator, and most of them are over empty ocean. The ones that are "owned" aren't going to be wasted on cheap satellites (although they might wind up being *large* high reliability monsters).

They have updated the software a few times. The well published bits include better error-correction and compression of data going in/out (a lot of research on that field has happened since 1977!). As you might expect, it has rather primitive computer[s?] and the code that gets uploaded would impress Mel himself[1]. NASA also tends to include code to work around failed bits (see Kepler for examples), so this helps with the maintenance. [1] https://en.wikipedia.org/wiki/The_Story_of_Mel

Note that in the case of Philae (project that bounced on a comet) and Osiris-rex (mission that ends with "landing"/grabbing a sample from an asteroid) are dealing with gravity from masses small enough to have CoG and CoM differ enough to influence maneuvering. It takes some pretty extreme sizes for this to be true, and typically only possible where the object producing the gravity in question is *tiny*.

Basically a design based on confusing cause and effect. Nearly all "magical" high efficiency rockets are SSTO, so the idea is to build an SSTO to somehow get that high efficiency. The other consideration are logistics issues, it *should* be somewhat cheaper to get an SSTO. That said, you better not have the issues of refurbishing the shuttle or you will never get the costs down to flying a falcon9 on a reused booster (even with a new first stage). SSTOs are likely to make sense with ISPs>>400s (or similar tech), or on planets with lower gravity wells than Earth. On Earth they are a sign of someone who doesn't know the first thing of rocket science.

One point of "progress" (in any field) will be the shear number of people and the amount of time/energy they can devote to that problem. A higher population implies a certain progress. A population that has a more "information based" economy implies that more people are available for similar problems, and likely build tools that "progress" your field (watch some of James Burke's Connections: most of a field's progress probably didn't come from that field). Another issue is how easily such progress can drop into the hands of those who need it (but might not have otherwise heard of it), certainly things like patents and IP laws can slow things down by 17 (or infinitely, with other forms of IP) and typically are dealbreakers when the guys in the lab aren't capable/willing to get into huge legal negotiations for "small" things. I was a bit shocked to look through some old Robert Heinlein books (I think it was Expanded Universe [1980ish] and point a "biggest breakthrough" would be google (described as "sorting through the mass of publication" and "the computer guys are leading the way" [don't ask how spot on his 50 year predictions were]). Being able to *get* those tools/information/whatever from other fields is critical, and a search engine makes it easy. To a certain extent, like generals fighting the last war, we seem to expect that the next great breakthrough will be just like the old. It rarely is. I also have to admit that not only do scientists so rarely "stand the old theory on its head", I have trouble imagining when that happened. Probably the best example would be how atomic theory and oxidation overturned the previous phlogiston-based chemistry. Even then, the phlogiston-based examples could be corrected by replacing "phlogiston" with "the absence of oxygen". It might have been a huge deal to the chemists, but anybody mixing chemicals would hardly notice the difference. Since then? Math still doesn't seem to notice that programming (or at least writing bug-free code before testing) is hard, and that all those proofs are a lot less certain that previously thought. I'd say that might be "stood on its head", but math isn't science, and the universe looks no different.

From the sound of the experiment containers, it should be even easier to spam science into labs. Just bring along a few dozen experiment containers and hit every biome on Minmus. Then take your dozen experiments * (sum of all data on Minmus) and feed them into a dozen science labs.

I suspect that to get 75% of the laptops, you will have to include chromebooks that can't play it at all. For anything much beyond a direct chromebook competitor, I've found that everything else works. Could you post the specs (especially the GPU) of something that doesn't work? The only one I've seen unacceptable is an old dual athlon (at least 10 years old) notebook. I really hope that the console controller [on the console game. Obviously doing it on a PC is trivial, assuming you have one] allows you to rotate and translate using left and right console controls: that is one of the bigger drawbacks of mouse and keyboard and the only reason I'd want to plug in a console controller.

I suspect they can do the backburn well enough or they would have stretched it during 1.1 (or possibly added room for SRBs). While I like the raptor engine even more, I suspect that is due to my lack of RO play (where I often run into needing moar thrust on my second stage. That doesn't seem to be a real issue in real life).

By "staying alive" I meant without Earth assistance. Don't count on it happening with a century or two. Also LEO is cheating and definitely not in "the asteroid belt", although I would think that for the asteroid belt you could be sufficiently far from the Sun that radiation shielding would no longer be that critical (and you would be sufficiently deep in an asteroid for enough of your life not to care). The asteroid belt has mining. Mars is being sold on "a second chance in case we cause human extinction on Earth" (and presumably Alpha Proxima would be "a place that will outlive the Sun by hundreds of billion years"). No idea what the point of an LEO colony would be (we still haven't found any of those wonderful microgravity manufacturing wonders).

Because while I might be able to get them back, they are basically just shells of steel. What I want back are the liquid-fuel 1[.5] stages. Work back from orbit: A Falcon 9 booster has to provide roughly 7-8km/s delta-v for a lightweight rocket (when the booster lands) or a 4-5km/s delta-v for a heavy one (where it doesn't). The upper stage of a Falcon Heavy will be able to provide even less delta-v (unless I missed something, it is the same as the Falcon 9) as it is designed for an even heavier load. This puts the middle stage as providing at least 6km/s of delta-v (and thus hurtling towards space at 6km/s, and have to boost back nearly the entire velocity). If you want to avoid that tremendous boostback, you need to add delta-v to the upper stage. Thus the SRBs on the upper stage. If you aren't worried about the boostback (and not only is Spacex apparently less worried about it, Blue Origins appears blase about it to with New Glenn), then don't bother (or put the SRBs on the first stage).

Reduced back burn on the recoverable stage. Boosters have so far been recovered after getting to 1-2km/s. A Falcon Heavy middle booster carrying a cargo beyond what a falcon 9 (unrecoverable, something like 28tons) is likely to require >6km/s. Coming down from those speeds have been iffy (although this middle booster has been specifically "stretched" for such a flight). Dropping the middle stage earlier and applying delta-v directly to the upper stage would be the whole point. I wouldn't be surprised if the economics dictate that spacex will simply throw away the center booster for heavier flights, and still be more economic than multiple flights with full recovery.

Oddly enough, had I worked on the Falcon Heavy, I would think long and hard about adding some COTS (commercial off the shelf) SRBs (solid rocket boosters) to the Falcon's upper (non-recoverable) stage. While reducing first stage mass is relatively unimportant, *recoverable* mass is even more of an issue. While sticking the SRBs on the upper stage might not help delta-vs as much (to be honest, they might help even if they slightly reduce delta-v), the point is that to reduce the delta-v coming from the middle booster. Since that booster needs to be recovered, keeping the delta-v low (and thus the boost-back low) is critical. Since neither the upper stage (nor the SRB boosters) would ever be recovered, they can have all the delta-v you need. Most of this can be found by either playing KSP or studying the rocket equation. Since the effective "dry mass" of the lower stages includes the entire upper stage, the Isp of the lower stages becomes far less critical, while the Isp of the upper stages becomes important (because the "dry mass" is only the dry mass of the upper stage + cargo). If you want higher mass fractions (and bigger cargoes) even the Isp becomes less important and you look to things like cheap power (which in KSP means kicker SRBs, but this isn't necessarily true about real life). Examples of getting big payloads into LKO (low kerbal orbit) on the cheap...

I guess the other question is "why SSTO"? Much of the reason for being 2,700 tons is that you don't care at all about mass and are willing to just keep scaling up. But what happens if you split it into two big dumb boosters? Since the engine shouldn't be a weight factor, I'm assuming that you wind up throwing away "leftover" fuel/oxidizer that isn't pressured (and has to be tossed to regain TWR). I've yet to see a reason to use SSTO with current Isp values, and would be shocked if this was any different.

Mars: probably easier to get a colony to stay alive, and hit self sustaining levels much sooner (i.e. in only a few centuries). Asteroid belt: Easier to get there and back (although I'd expect this means ion-style constant thrusting as there is no Oberth effect or aero assist during capture). Wildly easier to hit economic payback. The real kicker is why you would inhabit it: presumably to perform any tasks out there that robotics aren't up to (expect politics to be a huge part in determining this). I'd expect to at least start delivering raw material from the asteroids fairly early (should be easier then landing on mars), assuming you could find a sufficiently valuable asteroid that you could nudge into Earth orbit, mold into deorbitable capsules, and then recover (finding a place that will let you bombard them from space and still legally protect your "property" may be difficult).

I'm failing with google, but I'm pretty sure there is a "list of rules for rockets" that includes such things as "any program that includes a launch vehicle becomes a launch vehicle program". The idea being that designing a rocket capable of launching a spacecraft (into orbit and beyond) is so difficult and expensive that it overwhelms the entire program. Rocket programs typically find the most existing reliable rocket with the least cost (in that order) to provide the delta-v for the mass they need. I'd have to say the simple answer is to play KSP and figure it out. Note that you need not buy a copy to [mostly] answer the question, the game demo works reasonably well to learn the difficulties of getting to orbit (assuming that someone else supplies all the parts). If you did buy a copy, then installing the Realism Overhaul mods should allow getting the realism details much closer than stock KSP (hopefully you also downloaded KSP 1.1, getting Realism Overhaul to *all* work at 1.2 might take a while).

I'm pretty sure I remember Apollo-Soyuz (I was born before Apollo 11, but not by much). I also think I saw (on TV) Skylab's launch (and didn't they use Saturn1bs to get to it)? I always remember counting down "4...3.Ignitiion!.2..1..Liftoff" and was disappointed to learn that SaturnV (at least) didn't ignite at that time. I wonder if the Saturn1b used to go to Skylab matched that, or I simply was paying much attention (I may have grabbed it easily from a fictional launch as well, things like "stowaway to the Moon" were popular at that time).

Since base 2 can't be reduced, it would presumably count. But that is hardly the only way to represent numbers. Consider the Godel base system: describe a number by multiplying prime numbers together (I think his system used only single primes and thus was sparse, perhaps some sort of inductive means to describe the numbers of each prime...). Also using only ones and zeros hardly forces you to use binary. Binary is simply the most efficient means to code numbers with ones and zeros if you want to describe integers equally. But try looking up arithmetic coding (typically used in data compression, especially compared to Huffman compression): you can encode different bases (not just powers of 2) efficiently in binary (of course there will be leftovers for finite numbers when not using powers of 2). I really don't think this can be answered (unless you really pull off a base-e system. And then only if you manage to pull it off without "cheating" by using arithmetic compression). But I see no reason to believe that adding various powers of a single number times various coefficients is the only 'natural' number system.

Is mountain climbing inherently aggressive (it certainly was for early mountaineers. They planned campaigns to 'conquer' the mountain. More recent ones aren't so aggressive). How about drag racing (a sport mainly concerned with delta-v)? It certainly tends to be, but I'm sure plenty are more concerned with the time on their slip and the adrenalin rush and not at all in the car beside them (Jeb would love it). I'd say not. The shear *cost* of spaceflight on Earth simply puts it out the range of all but a few, and of the few the ones with the most interest (mostly in ICBMs) are military.

Certainly way farther than spacex ever did with grasshopper. And I was astonished to find that the first two falcon attempts involved parachutes (although I was certainly one of the many advocates of starting with parachute landings), so spacex had a much further path than I thought trying to use a powered landing on a booster designed to land with parachutes. But it is certainly a lot more delta-v. I wonder how much of a fraction the booster is going to deliver? Falcon 9 delivers 1-2km/s, but appears to be designed around the "expendable mode" (which provides more like 6km/s). If they want to "go big", I'd expect a pretty significant boostback (more than spacex's ~1km/s).

It sounds like any high-coverage data from such satellites would work, such that you could remove something like 10% of your data as a control system (and presumably break that up into sub-groups: training with less than the full group and comparing it to the the sub groups, then taking the best of those and comparing it to the "real" controls). What I'd really want to do with aerospace and neural nets is to hack up a kOS routine that would land a booster spacex-style (or possibly simply get into orbit). I strongly suspect that C# and kOS just aren't up to the challenge (of building/training/running a neural-net) and that trying to find another rocket simulator would be prohibitive.

Bigger might not be harder, but it likely makes learning more expensive. The delta-v on the booster stage is almost certainly far beyond anything the New Shepard had to deal with. Presumably they are launching from the cape, have they announced if they are landing on land or at sea? New Shepard only had to deal with vertical velocity and barely controlled horizontal in either direction. New Glenn will have to control in at least one horizontal direction as well (and hit a target while moving fast, they hit the earlier target dead on but with little movement away from it). The economics make things more interesting. Spacex put 19 boosters in the ocean before managing to recover their first one (and they still haven't launched a landed booster). If recovering boosters is the goal, I'd expect to go smaller in the idea that you can learn faster with cheaper boosters. If launching large amounts of tonnage into space is the key, bigger is better and keep dumping the boosters until you learn how to land them. Spacex almost certainly built the falcon 9 that size to allow sufficient cargo to pay for the landing attempts (throwing more tonnage to space is pointless if you don't have the customers). If Blue Origin is less concerned with customers, then bigger is better. One thing you shouldn't forget about are launch costs. DC-X made huge strides in reducing launch costs and these probably carried over into New Shepard. I suspect that once you start launching at the cape, NASA will inflict their customary cost schedule on you, and expect to pay nearly the cost of a Falcon 9 booster just to launch into orbit. This means that once you manage to recover a big booster, your relative costs (costs/ton) are going to go way down (assuming the launch costs are mostly similar to go into orbit. I suspect manned costs will balloon again, but that is unavoidable). To be honest, I suspect that the choice of size was entirely driven by ego. Given that all sizes were effectively on the table, Bezos went with the big one. It would take a lot of smaller rockets to justify an intermediate stage if you are really interested in the big cargoes, but if you wind up dumping 19 into the drink then a smaller prototype rocket makes sense (even if you don't have paying customers due to contracts with ULA).

I was always impressed by Escape Dynamics' work toward microwave powered flight (using a ground-based heating device to heat H2 in the rocket. Instant 800 ISP). Unfortunately, the microwave on the ground would be impossibly expensive (I doubt the company could really afford the vehicle, but the engine was the killer). Last I checked, the Navy was developing a laser with the wattage needed by Escape Dynamics (note that a laser weapon likely needs full power for much less than a second, while the launch vehicle needs it the entire launch (or at least until it has reached a sufficient apoapsis and preferably an insertion burn (although that is likely through too much atmosphere). Who knows, maybe such weapon research will find its way into a similar launcher.

Not with government contracts. You deliver what the thing says, and absolutely *zero* more. Don't be surprised if it is a felony to deliver more (you certainly have to charge every hour worked, that gets driven into the lowly workers). This might be slightly different in other branches (an old Army hand I worked with was doing quite well as an Army contractor after crashing and burning with Lockheed and the Navy*, but it was Lockheed I was most familiar with and if the contract doesn't include oxygen, don't expect to breathe). In the unlikely event that ULA had a commercial cargo that wasn't delivered by lobbyists then they would be careful not to screw them over, but government is a different beast entirely. * The ex-Army officer kept supplying "support" after it had be removed from the contract. Even I, who knew next to nothing about contract management, knew you couldn't do that in our business. He took one of our better techs with him and was last heard doing fine (so I suspect this isn't true everywhere, but I wouldn't expect to breathe if I didn't pay Lockheed for oxygen).

There's little meat for this thread until spacex fixes the launch pad. And they can't start fixing it until every last scrap of evidence is examined. Maybe we can at least talk about the cause of the explosion while they fix the pad (and publish scraps of potential causes).