Neat. Yeah, it is a bit of a poke in the eye to NG. Still can’t wait to see Block 5–so far they’re just going for a single reuse each of Block 3/4 boosters. Block 5 should allow 10+ reflights.
In other SpaceX news, they just showed off a picture of some new BFR tooling. They use a robot to wind carbon fiber around the cylinder, which then forms the main body of the second stage. It’s a big sucker (see Model 3 for scale).
The Bangabandhu-1 launch, later this month, should be the first “Block 5” flight:
Out of curiosity, is the ground support equipment changing for the block 5?
Nothing has left 39A since FH, and we know that future FH will be block 5, so Elon’s minions could be modifying the TEL to reflect the new normal, while 40 and Vandyland get rid of the older boosters. The next thing from 39A is, you guessed it, a block 5.
I wonder what Musk would charge NASA to rush job a high speed comm bird to Mars orbit? The bird they are planning to use for sample return relay work is today on year 12 of the 5 year plan, and they are talking mid to late 2020’s for the return mission. Sounds like worse planning than normal.
I think that article is a piece intended to represent the Boeing-Lockheed joint venture ULA, not excuse NG.
Besides, NG seems (to my eyes at least) to be toward the bottom of his most favored contractor list, based on very critical articles like this:
I guess we can conclude that Boeing “contributes” more to Loren’s Lexington Institute than NG, who contributes more than SpaceX (who contributes nothing).
At any rate, the article certainly doesn’t argue against Loren being a complete tool. 1.5 years later, has Trump reversed any of the “decay” that allegedly occurred under Obama?
I don’t believe that NASA wants to rush anything. They want equipment to work, and astronauts to return to home.
I’m pretty sure SpaceX wants those last two things as well. Personally, I’m glad they’re a bit more concerned with pushing the envelope and making progress than NASA (but that’s pretty easy to say when there’s no chance I’ll be on their first manned mission -or even their first hundred probably).
NASA needs to have a more risk-accepting culture. That doesn’t mean putting astronaut lives on the line, but something like a Mars comms satellite certainly belongs in this category. It is cheaper to build fast and iterate on failure than it is to be an absolute perfectionist, particularly now that SpaceX has cut the price of launches so much. SpaceX demonstrated this development model with their reusability program. It suffered many failures but was nevertheless far cheaper and resulted in a better product than traditional style programs.
When NASA first tried to implement the “Faster, Better, Cheaper” (FBC) paradigm under the leadership of Dan Goldin they got failures such as the Mars Climate Orbiter and the Mars Polar Lander. Given the relatively fixed nature of the NASA budget and while that is largely discretionary spending that is allocated to programs on an adminstrative level, the agency has to appeal to politicians who hold particular sway, mission failures can be highly problematic for the agency and particularly for the managing center. FBC and the failures resulting from it arguably had an impact on reducing the budget allocations and staffing at JPL, resulting in early retirements and layoffs that have left it berift of the experience it developed during Voyager and the development of Cassini-Huygens.
There are a few misconceptions here, the biggest of which is that NASA has any real choice about their risk posture or accepting failures as part of normal business. Every significant failure or mission loss comes with finger-pointing and recrimination regardless of some of the past amazing successes such as Apollo, Voyager, Cassini-Huygens, Hubble, or the Mars Science Laboratory. The agency is currently being villified for delays and cost growth in the James Webb Space Telescope even though those delays and cost growth were specifically due to having to stretch out the schedule and reduce annual budget to pay for unexpected increases in the crewed space program. NASA is risk adverse in their large programs and missions because their failures are not tolerated. That being said, they do innovation on smaller, under-the-radar programs including on the vehicle recovery/reusability program (Delta Clipper Advanced) that predated SpaceX Falcon 9 Stage 1 autnomous landing by more than a couple of decades.
The build cheap and test model works well for things that are quick to build and test. Interplanetary missions are not cheap nor quick, and because the hardware may be exposed to extremes of thermal and radiation environments extensive testing is necessary to expose design and workmanship problems. Early in its existence as a NASA center the Jet Populsion Laboratory (JPL) embarked on a program of quick and cheap remote probes in support of the Lunar landing effort (the Ranger program) of which the early failures were almost the undoing of JPL. If NASA would be funded to perform a series of say, Mars landings or outer planets missions using common hardware and modular software, they could potentially perform a test-fail-improve cycle, but at the current rate of funding the agency gets about one Mars mission and one outer planets mission per decade. NASA is not an efficient organization on many levels, but that is a result of how the agency is composed and funded, and to an extent the type of missions it is chartered to perform.
It is unknown just how much SpaceX has invested into their development; as a private company they do not have to make their finances public, and it is unclear that even they have a full accounting of invested and operating costs. They have been able to do the devleopment at lower cost by using private venture capital and bidding on FAR Part 12 government contracts rather than having to provide test and manufacturing data or submit to full mission assurance provisions of the traditional FAR Part 15 contracts like the previous EELV contractors, but that also means that as a payloader you have to have more faith than insight into the vehicle capabilities, which is fine when it is operating well within demonstrated capability but can be problematic when mission requirements press the boundaries of prior experience, e.g. performing long coasts between restarts or launches to new azimuths.
That SpaceX has been successful in both demonstrating good functional reliability (despite a couple of failures) and new capabilities does speak to benefits in following their key design parameters marketspace vice traditional requirements-based systems engineering (although the fiscal advantages to reusibility have yet to be assuredly demonstrated) but applying that approach to missions that require many months or years before being able to evaluate success rather than just a successful launch and recovery is another situation entirely.
Stranger
Right. I’ve pointed this out in the past as well; that NASA’s low risk tolerance is a function of their funding model–i.e., Congress and the Senate.
I wish there were a solution. I don’t have one. NASA has backdoored this a bit via their COTS and CRS programs, which were (are) vital for SpaceX and allowed them to succeed in areas that NASA themselves could not. They managed to fly enough under the radar to succeed before it really attracted people’s attention. I would like to see them apply the same partnership model to other programs but this may have been a one-shot deal.
I agree that the current planetary exploration model is not really compatible with rapid iterative development, but this is not inherent to the idea of planetary exploration.
For instance, significant Mars exploration–even before any human presence–really requires a network of communication and positioning satellites around Mars. As with Earth constellations like GPS or Iridium, the satellites should be small, easily replaced, and be incrementally upgraded. Because it’s a constellation, it can handle some degree of failure.
We could have a program of sending a small batch of these satellites to Mars every two years. If a batch failed early or had some other issue–big deal, it’s only a few hundred million dollars, and they’ll figure out what went wrong before the next batch. The satellites themselves should cost several million each, not be irreplacable billion-dollar one-offs.
Of course, there are no serious proposals along these lines, so instead NASA spends billions on dead-end systems like SLS. I understand why, but I don’t have to like it.
At some point I may start a thread about this topic: what is it with the obsession with Shuttle derived vehicles? SLS is only the latest in many previous efforts. I sometimes see Kerbal Space Program players get chided about “rockets aren’t like LEGOs”, but whoever keeps coming up with these systems that plug together N-segment solid boosters, stretched/modified external tanks, RS-25 variants, and so on apparently didn’t get the memo. It can’t just be senator influence–or can it?
Yes, it can. That, plus the lack of funding for R&D to come up with something really new. But I really think the majority of the reason is pork and cronyism in Congress.
When Apollo first started, NASA intentionally made sure that large important facilities were built in as many states as possible, precisely to gain political support to maintain its funding because they tied the size of NASA to many jobs in many States.
I guess they just forgot that their dependence on those votes cuts both ways and gives those Senators a lot of power over appropriations and which programs will get funded.
In my line of work (systems engineering), I see this a lot: “reuse engineering”. Always sold as cost-reduction and risk-reduction. (In other words, beneficial to someone who’s risk-averse and cost-sensitive.)
True innovation is rare in NASA; the burst of creativity in the NACA and early Space Race era has carried us through to today, and it’s “good enough”. Even scientists aren’t particularly innovative there. (I’ve worked NASA programs until very recently, so I saw this first-hand.)
Innovation has to come from someone not invested in the status quo. Hence, SpaceX and the ilk.
I don’t have any direct knowledge, but this reddit post certainly seemed to be thinking along the same lines as you:
Perhaps I was vague, I was thinking that if NASA needed to get a replacement relay bird to Mars because the grey-haired one the have now was near death, that SpaceX might have a booster ready and waiting, verses the other launch providers who might just then start bending Al. Having a working production line can make short work of wait times, when cash can be waived.
Second thought: How long before someone starts mining the spacecraft graveyard in the Atlantic for sunk boosters and rare earths? After 50 years, the concentration has got to be better than raw ore now…
You are making a real apples and oranges comparison here. First of all, Iridium is operating a worldwide satellite consumer communications service. The company maintains a constellation not for reliability per se but for near global coverage, and while they do obtain certain production economies of scale through the quantity of satellites and launches, the individual satellites don’t just cost “a few million dollars” but over US$40M for each Iridium NEXT satellite not including launch costs. The satellites themselves are located in Low Earth Orbit (approximately 781 km circular by 86.4°) where they are largely protected against solar particle radiation by the Earth’s magnetosphere, communicate among themselves on the Ka band and to users via L band frequencies.
By comparison, communication satellites orbiting Mars and relaying directly to the NASA Deep Space Network would have to communicate in the X-band or S-band frequencies at much higher power, would have to survive solar particle radiation events without any outside protection (Mars has no magnetosphere), and their deployment and successful function won’t be known until several months after launch. While there are multiple potential windows for launching into LEO every day from virtually any launch facility, there is a fairly narrow window for a pseudo-Hohmann transfer once every 26 months, so ensuring that all aspects of launch, transplanetary injection, capture, and final deployment as well as successful operation of the satellite on-orbit is critical not only in terms of not wasting money on failed deployment but also in supporting other programs which would depend on the capability, hence the high degree of testing and oversight to ensure maximum potential for success (and the illustration of how an FBC-type failure can undermine future missions).
“A few hundred million dollars,” is the entire operating budget for supporting and managing a decade interplanetary missions for a center like JPL or Goddard (not accounting for the development and launch costs of the actual vehicle), and the loss of a mission that is revealed to be for a stupid error such as selecting a fastener prone to corrosion or not performing a a good stating and recovery analysis pretty much guarantees that heads will roll and future budgets will be scruitinized, hence the aversion to adopting a risk-accepting posture. Despite the broad (and not entirely inaccurate) characterization of NASA as a wasteful bureaucracy that accomplishes little for the dollar, the interplanetary missions are often run on a relative shoestring budget and provide enormous scientific yield for the dollar investment compared to, say, the International Space Station or the Apollo Lunar Exploration programs. The mission management is very focused on staying within budgets because unlike the prestigious crewed space program they cannot go back and demand more budget, and when cost growth occurs, it is often as not due to cuts in annual budgets that stretch a program out and result in the underlying fixed and administrative costs growing to a larger portion of the overall cost, which is the case with the JWST.
As far as a model of how space exploration should be run, Iridium has built some amount of loss and premature failure into their business model which lets them absorb the costs of a launch failure or non-operation of a satellite without turning into a major crisis, but the company is also famous for also being the largest bankruptcy filling in the United States at the time (2002) despite being technically viable and mostly within estimated deployment and operating cost, and became profitable only because of US Department of Defense users needing it for the Global War on Terrorism and the wars in Afghanistan and Iraq, so I’m not sure pointing toward Iridium as any kind of model for guaranteed success is justified by evidence even on a commercial profit basis. SpaceX also was on the edge of bankruptcy prior to a string of successes with the Falcon 9, and it is unclear whether they are clearing a profit with current operations much less recouping the original investment in design and production.
The entire charter of NASA is not to generate any kind of direct profit but rather to perform the space exploration and Earth surveillance missions it is assigned by the executive branch and as funded by Congress. Therefore, its focus is on maximizing the potential for scientific yield rather than minimizing individual mission costs. If NASA were funded for, say, five outer planet missions per decade, they could probably drop the per-mission costs by half from 6-8 US$B per to something like 3-4 US$B per and take the risks that an individual mission might fail to meet all objectives; unfortunately, the next NRC Planetary Science Decadal survey is likely going to recommend one Saturn mission and one or two Mars rover/sample return missions, and that will be it for planetary science for the foreseeable future. There isn’t really a way to optimize or reduce costs at those mission rates, nor is it feasible within the current infrastructure to perform missions with cheaper, lower powered probes or landers, so it is crucial for NASA to make sure their missions are as free of risk and have the highest potential for scientific return as possible.
As it happens, I’ve worked on a study for a communication and telemetry infrastructure which would support lower cost missions using smaller or multiple deployment spacecraft and would be crucial for the expansion of interplanetary exploration beyond the already strapped and obsolescent terrestrial Deep Space Network (DSN). Despite the pitch for a constellation of satellites around Mars, the bottleneck for Mars or other interplanetary missions is not the on-orbit communications system but rather communication back to Earth from missions going in different directions and competing for limited data throughput. Our recommendation for a 3-6 unit array of solar orbiting satellites between Venus and Earth orbit which would also double as observatories to search for potentially hazardous objects (PHO) that could not be seen from ground-based telescopes was essentially rejected out of hand because although the costs were comparatively modest and would represent a vast expansion in data throughput and reliaiblity compared to the DSN, the costs were not factored into any long term budget and could not be paid for under individual mission funding (which is largely the way the DSN is funded), and also provided more capability than was projected to be needed in the foreseeable future.
But regardless, “rapid iterative development” isn’t compatible with planetary exploration not because of any sort of policy position or costs but because the time and distance involved in planetary exploration and the baseline costs of supporting such missions don’t lend themselves to the fly-break-fix-fly approach. If propulsion technology were better, or there were a complete infrastructure to support in-space manufacture and resource extraction which would make it viable to launch and support concurrent and overlapping successive missions that would be different, but at the current state of the art and the infrastructure that exists today the notion of cutting planetary exploration costs by orders of magnitude is no more feasible than Elon Musk’s original claim about how the Falcon 9 would fly payloads to orbit for a few million dollars on a daily basis. That’s a nice goal to get there, but just tossing up a PowerPoint presentation with flashy animations does not make it so.
The rationale behind using Space Transportation System (“Shuttle”) derived vehicies is that it would leverage off of the existing manufacturing, integration, software, and mission support infrastructure but cut away the things that made Shuttle missions so expensive and problematic, e.g. crewed flights, the heavy wing structure with delicate tiles, et cetera. It made sense back when the Shuttle was expected to fly one mission a month and an assumed 2-3 week turnaround time. In the 2000s and the post-Shuttle era, it is more about maintaining the current contractor workforce and minimizing design and qualification efforts, although (predictably) making substantial changes to the configuration versus the basic STS configuration requires substantial redesign and full qualification, and the once-reasonable costs to produce the segmented large Shuttle Rocket Boosters have increased substantially due to low production and much higher costs of constituants. From a reliability standpoint it makes sense to stick with the SRBs which, after redesign have flown more than 200 individual flights (two per launch) without significant anomaly, despite often ill-informed criticisms of the design or use of solid rocket motors, but from a cost and expansion of future capability standpoint, the SRBs are an evolutionary dead end despite the convenience and reliability they offer.
NASA was mandated to develop the SLS (often caustically described by detrators as the “Senate Launch System”) and from what I’ve seen there isn’t even a lot of internal enthusiasm for it despite having been largely derived from the unsolicited DIRECT proposal and Jupiter rocket concept advocated by NASA managers and engineers working outside of the Constellation program. The cost and flight rate means that it will provide the astronaut corps with very limited opportunties for flights, and the lack of an essential mission gives the entire effort some amount of question whether it will ever fly more than a handful of flights or perform any specific programmatic goal.
I would take your criticism one step further and ask, “What is the obsession with tall and slender cylindrical multistage rockets?” Even companies that seek to be innovative in their approaches to space launch have largely stuck with the essential planform of WWII-era sounding rockets despite the inherent problems with them. The arguments for sticking with this system are largely based upon manufacturing limitations and the logistics of transporting rockets over the road or barge, but if the eventual plan is to launch and recover boosters in powered flight there is little reason to stick with a roadable diameter or the conventional manufacting flow, and there are substantial operational and mass optimization advantages to a more squat planform. But of course this represents a large divergence from current practice and the one time this was seriously proposed (the Chrysler SERV/MURP) it was rejected essentially out of hand, not just because of the difference from all other proposals or the required demonstration of novel technologies but also because it was not well-integrated into the existing aerospace and defense contractor manufacturing complex. The next generation of large space launch vehicles shouldn’t look anything like either Shuttle-derived vehicle or the Saturn V, because while those were the ostensible best that could be done with the tools and knowledge of the time, they are certainly not the best thing that could be done with modern design analysis tools and current knowledge.
Stranger
Agreed. It wasn’t intended as a very close analogy–but I think the principles behind a real constellation, with upgrades and redundancy built into the system, are sound even for Mars.
Satellites are getting cheaper. As you point out, the original Iridium was not a financial success until the DoD decided it would come in handy. But the Iridium NEXT birds are both substantially cheaper and far more capable than the originals, and cost significantly less to launch. It seems likely that Iridium NEXT will be a success.
In any case, my point was not to demonstrate that a Martian communications network is somehow financially comparable to various Earth networks. Just that we already see the movement from huge, expensive, monolithic systems to smaller, cheaper, and rapidly iterated constellation-type systems. It is not the right solution for everything but I think there is a distinct pattern here and there is no reason it can’t be applied outside of Earth as well.
This is unfortunate. Even though I think they could do far more even at current funding rates, planetary science is nevertheless one place that NASA excels at. I agree that additional funding would have an outsized effect both in efficiency and in the ability to take more risks.
A constellation can serve the entire planet without blackouts, regardless of where the landed vehicles are. And of course, what I have in mind is not just one or two rovers; I would like to see dozens of landed vehicles, including rovers, drilling units, sample return platforms, and ultimately stuff related to human landing prep (“bulldozers” to prep landing pads, ISRU processors, water extractors, etc.).
This is wishful thinking of course. Maybe someday.
This sounds like a fantastic proposal, and something like that would have to be a part of any other serious interplanetary network upgrade. A shame that there was no funding model. I like the idea of an integrated NEO detection system.
Well, I understand the superficial argument. When I was 14, I would have agreed–I was reading about the Shuttle-C in those days and thought it sounded cool.
I just find it crazy that there’s such a disconnect between the people writing the checks and the people that really know this stuff. Someone, somewhere must have pointed out that development reuse like this is the exact opposite of savings: not only do you have to redesign everything essentially from scratch, but you have to do so with design constraints that you would not have with a cleansheet design. It’s the worst of all worlds.
Of course, it can get worse than SLS/Constellation/DIRECT/NLS-like vehicles. The worst idea was probably the SRB-X proposal, which was to have up to three SRBs, with one as a “center core”. Bleh.
It’s a fair point. SpaceX is getting away from slender rockets a tad with their BFR, which has only a 5:1 aspect ratio (compared to the Falcon 9’s 19:1 ratio).
There is something to be said about cylindrical production and its benefits. And while important, absolute tankage efficiency is not quite as important on multi-stage rockets as on an SSTO, particularly given that methane and not LH2 appears to be the next-generation fuel of choice. LH2’s low volumetric efficiency demands optimal tankage–the constraints are less with other fuels. Vehicles like the SERV, DC-X, and VentureStar had to have a chubby aspect ratio for this reason.
Nevertheless, I do anticipate a move to a squat aspect ratio with a hypothetical next-next-gen ultraheavy rocket (1000+ tons to orbit). The reason is simple: rockets are roughly as tall as they can get. A rocket engine has to lift the column of propellant above it, and they can only be packed so densely on the first stage, and we are probably not going to see huge leaps in chamber pressure. You can go a bit taller with a conical shape, but to increase volume beyond that requires a rocket that is simply fat. Fortunately, there is not much downside to doing so–it does not make the aerodynamic situation any worse, and there are some tankage efficiencies.
Correction: the BFR aspect ratio is more like 11:1. Still significantly fatter than the F9.
But who’s buying the argument? An undergrad engineer of almost any variety can see that you can’t just take a Shuttle ET and put a rocket on top of it without essentially redesigning the whole thing. Loads are transferred through internal structures, and these are going to be totally different between the two cases. I get the desire to reuse stuff. But even a cursory look shows what a bad idea it is.
Are the senators willfully ignorant? Have people at NASA been telling tall tales about this stuff? Hell, has someone been reading too many dodgy sci-fi novels (I remember a series, by Stephen Baxter I think, where a significant plot point was throwing together a bunch of scrap Shuttle parts into some kind of interplanetary spaceship).
The Mayor of Los Angeles tweeted out that SpaceX will start BFR production in the port of LA: https://twitter.com/MayorOfLA/status/985931907970908161
It’s not quite as bad as that. The External Tank (ET) is designed to carry loads from the Solid Rocket Boosters (SRB) through the backbone to the Orbiter Vehicle (OV). While changing the ET to a forward thrust structure requires redesign of the forward interface, the main load path is through the tank walls which, having been designed to accept bending loads from the OV thrust and mass, are more than sufficient to accept CEV/Orion Command and Service module direct axial inertial loads. The bigger challenge is actually in the aft end where a completely new thrust structure had to be designed to accept the RS-25 engines, and oh, by the way, putting the engine controller into an environment it was never designed to endure, which caused some major problems in qualification.
The real point of the SLS was to reuse existing support equipment and production facilities, and while the impetus for this was to maintain jobs in those critical districts, it would have made some kind of sense if a) most of that production wasn’t already dramatically scaled back with many of the workers and engineers retired or moved on, b) there was a clear design path for optimal value in reuse rather than just sticking with existing features to minimize non-recurring engineering costs, and c) there was a specific mission need for the vehicle at a high enough flight rate to justify several flights a year rather than jamming four astronauts elbow-to-asshole into a capsule and sending them to the International Space Station (ISS) or on the now-cancelled Asteroid Redirect Mission (ARM). (BTW, nobody in the astronaut corps seems very happy with the Orion capsule, and specifically the decision to go for an ocean landing mode without direct recovery features.)
The SLS is, to put it short, an example of a good concept executed in almost the worst possible way. If work had begun on SLS in, say, the 2000 timeframe with the intent of using it as a heavy lift vehicle and then maturing it into an evolutionary replacement for the Shuttle using a smaller lifting body personnel vehicle while maintaining the SRB and ET manufacturing capability, it would have made somewhat more sense. As a rocket without a real purpose or mission other than to have something to fly and stuck with design choices made forty years ago, it is a predictable mess.
Stranger