Treating Space Access As A Consumer Product
SpaceX’s final Falcon 9 design coming this year, two Falcon Heavy launches next year, SpaceNews
“We are flying Block 3s right now,” Shotwell said. “Block 4s start flying shortly, and then Block 5 at the end of this year. We definitely have gotten better [at] more smooth introducing of change. You don’t see the big impacts to production we’ve had before when we’ve changed vehicle designs.” SpaceX Chief Executive Elon Musk had previously stated in an online question-and-answer session in October 2016 that Block 5 would be the final upgrade to the Falcon 9 design. The Falcon 9 Block 5 is expected to be far more reusable than the Block 3. Shotwell said a Block 5 booster could relaunch ” a dozen or so times.” The Block 3, by comparison, has an estimated life of two or three missions. Shotwell said the Block 5 version of the Falcon 9 won’t need refurbishing, but will mainly undergo inspections prior to launch, streamlining the process compared to the first reused boosters.”
SpaceX will try for third Falcon 9 launch in less than two weeks, SpaceflightNow
“After back-to-back launches last weekend, SpaceX could launch its next Falcon 9 mission as soon as Sunday from NASA’s Kennedy Space Center in Florida with a high-power Intelsat communications satellite. Liftoff Sunday will hinge on the ability of SpaceX’s launch team to prepare KSC’s launch pad 39A for another flight after the successful June 23 blastoff of a Falcon 9 booster with the first Bulgarian-owned communications satellite.”
SpaceX Launches Two Missions in Two Days (with videos), SpaceRef
“SpaceX has upped the ante when it comes to reusable launch systems. Within 49 hours the company launched and recovered two Falcon 9 first stages while putting it’s customer’s payloads safely into orbit.”
Keith’s note: Imagine if NASA adopted the same incremental product upgrading approach – one based on operational and customer experience/feedback – with forward and backward compatibility such as what SpaceX has adopted. Imagine also that this ability to improve a product is a feature inherent to the original product design and not an afterthought. Imagine all you want. NASA is incapable of doing this or even grasping how to do it – but there were faint glimmers of it during development of the Apollo/Saturn program. Cost and performance benefits aside, having a flexible space launch infrastructure like SpaceX has (I’m sure Blue Origin is no different) with inherent design resilience launched in a steady cadence is how space will be best utilized – not by using monster rockets that NASA cannot even afford to fly more than once every year or so.
” not by using monster rockets that NASA cannot even afford to fly more than once every year or so.”
The big irony here is that if all goes even part way to plan (and I personally believe it will) Space-X will have a launch system with “Monster Rockets” far bigger and more capable than SLS will ever be even in it’s final configuration and these will be a tiny fraction of the price per launch and reuseable to boot, NASA has always fallen foul of a catch 22 situation were what they build costs so much to operate that there is then never the budget to refine and upgrade in such a way as to lower the costs and improve the overall useability of the system and hence they bounce helplessly from one overblown and less than optimal spaceflight system to another like a nighhmare, never ending game of pinball.
Hindsight is always 20/20 but imagine where we might be if the Saturn V was evolved rather than scrapped. Would the notional designs for a fly-back first stage ever become a reality? Would re-usable F-1 engines have become “a thing”? Makes for an entertaining thought exercise.
IMO, F9 style stage returns weren’t going to happen until precision GPS and compact, relatively inexpensive voting computers arrived.
F-1 could absolutely have been reused. Copies of the F-1 engine were fired multiple times for longer than “full duration” on the test stand during qualification testing.
Or as I’ve heard it said, “there is nothing fundamental about a liquid fueled rocket engine that makes it expendable”.
Well, some of them are fundamentally expendable. For deep space applications, it isn’t uncommon to have one-time-only valves, sometimes opened or closed with pyros. That makes reuse a bit hard. But I agree that this sort of design isn’t mandatory for liquid fueled engines.
Yes, liquids are easier to reuse if you don’t include single use items in the design like pyro-valves and chemical based igniters. But I was talking more about fundamentals, not design details.
Compare that to a solid rocket booster. Yes you can reuse a solid rocket booster, but doing so is decidedly non-trivial to the point of “remanufacturing” not reusing. Due to several characteristics of solids, it’s just fundamentally a p.i.t.a. to reuse them. Cleaning out the casing (usually to bare case material) is a pain in and of itself.
Anyone who has seen the hydrolasing shop at KSC (basically a 7000 psi pressure washer) where the residual fuel and insulation were removed from the SRBs would agree with you. One of the SRB recovery divers told me the Shuttle boosters were not so much reusable as salvagable. Nevertheless no one has replaced pyrotechnic devices with pneumatic latches and actuators with the determination of Elon Musk. Except maybe Jeff Bezos. The Falcon has ingenious extendable landing gear, but the Shepherd has landing gear that can also retract. Watching them cycle (extend and retract) seconds before launch was thrilling.
Supposedly Falcon 9 Block 5 (due to start flying this year) will have redesigned landing gear that can retract. Presumably this means they’ve done away with the crush core inside the legs. When the cores crush differently for different legs, this results in the stage leaning after landing. I would hope that the “lean” has also been addressed with this redesign.
NASA is a very mission-oriented organization. The whole concept of incremental product upgrades doesn’t really fit. It implies a product, not a mission, and customers or users.
The usual process within NASA is to have a mission, with specific goals. Then you write down requirements which are necessary to achieve the goals. Then you design and build hardware which satisfies those requirements. The whole concept of an upgrade implies someone failed to write down all the requirements. Building in the ability to easily upgrade is, in effect, expending resources (time, money, effort, whatever) on exceeding requirements. It shouldn’t be necessary if you’ve done your job in correctly defining the requirements.
Even SLS, with its Block 1, 1B and 2 configurations, isn’t really an evolving design. The requirements, at least the level 1 and probably the level 2 ones, are already written. It isn’t as if they are holding off and waiting to define Block 2 based on experience with Block 1. It’s just that there are some goals can be achieved without all of the Block 2 requirements, and NASA is using Block 1 to achieve them sooner rather than later.
I’m not saying I agree with this approach or way of thinking. I’m just describing it and saying it doesn’t fit with incrementally evolving a design based on experience and customer feedback.
Yes, if NASA was organized and sponsored correctly for the future, we would have already been on version D of the Space Shuttle by now. Oh well.
Actually it is NASA’s bread and butter. TRL – technology readiness level. They do this as a matter of routine.
https://en.wikipedia.org/wi…
That’s not how I see it. I was talking about using flight experience. That’s pretty much TRL 9 by definition. I guess you could call it very efficient cycling of TRL 9 (flight) experience into advancing hardware from TRL 8 to 9. Usually, I think requirements for what the TRL 9 hardware will eventually be like are written when the hardware is down at TRL 5 or 6.
One of the problems I have always had with NASA is technology will be started then they start to mature it… then a new congress / president / administrator comes in .. shelves it for new stuff. And even when they do finish they still just end up shelving a lot anyway like the new J2 engines.
I would like that NASA push the TRL and they automatically shovel it into the private sector so NASA can by COTS (commercial off the shelf)
NASA’s “mission” focus is much of its problem. NASA considers flights to be “missions”, rather than components of a broader mission. The Shuttle flew 135 times. That’s not “a” mission, nor a series of missions. That’s a tool to perform tasks, and tools are developed incrementally/iteratively. Likewise on your side of the shop, probes are tools to do science, and science is an inherently iterative process. Instruments are the products, PI’s are the customers.
IMO, your later reply about TRL is a perfect example of that cultural failure. There’s no NASA nomenclature to even describe incremental/iterative development beyond “flight hardware”, you had difficulty even describing what you meant using that terminology.
And while some people at NASA intuitively see the point of iterative development of flight hardware, I’ve spoken (online) with people from NASA or the major contractors who seem to have a deep cultural blindness to the concept. They literally can’t understand what you are saying. And IMO it’s reflected in the culture, the practices, the budget requests, the terms-of-contracts for suppliers, in the very language used throughout the agency. It worries me. I don’t see a way out.
—
Even COTS (Commercial Orbital Transportation Services): While I’m a huge fanboi, IMO it carries exactly the same bias as I’m talking about. As COTS it was a tech development contract, a way of determining if the technology was at TRL9 with minimal cost/risk to the agency. But once proven, it immediately became a fixed CRS supply contract, with no further development. There was no continuing COTS program(s) with broader goals.
CCDev only overlaps through SpaceX/Dragon. Without that coincidence, they’d be utterly independent programs, with no overlap in suppliers or technology. And CC will go through the same fixed two stages. Development, then supply of a finished and finalised product; with no COTS-like further development, to push the hardware to the next level.
And even if someone at NASA is able to sneak through a COTS-like program to build a CRS-like resupply capacity for the proposed DSG lunar stationette, it will be a completely separate program. And DSG itself won’t be part of an iterative development (for eg starting with ISS expansion), in spite of the multiple vendors capable of doing such work.
Well, I’m not going to argue about the distinction between an ongoing program and a “mission.” That’s fairly central to my argument in favor of higher failure rates. (Risk of failure is a bad thing if you look at missions one by one. If you look at a program as a whole, where lower cost means higher flight rates, accepting the risk of a specific mission failing can actually increase the number of successes.)
To some extent, COTS-like programs do have the potential for evolutionary development. I assume NASA won’t mind Dragon resupply missions flying on a Falcon 9, block 5, as opposed to the block 1 they originally accepted. Since the improvements increase profits for SpaceX, curtesy of the fixed-price CRS contract, they have an incentive to make improvements.
And one place where I do see some real, rapid, evolutionary progress in flight hardware is in CubeSats and small satellites. The other COTS (commercial, off-the-shelf) is really working out well. You can get things like star trackers, thrusters, solar panels, etc. with really fantastic performance for the resources (mass, power, cost) required. Even as mission-focused an institution as JPL has a deep space transponder for CubeSats which is on version three (well, 2.1) and 2.1 is much nicer than 1.0. And that’s in under five years.
I don’t think that’s directly related to CubeSats being small, or to the fact that they are 10 cm x 10 cm x 10 cm x N. It’s because they are cheap enough for frequent flights and have a standard size/shape (regardless of what the standard is.) The frequent flight rates mean there is a market for production (not one-off) components, and an incentive for the suppliers to gradually improve their designs. And a standard form factor (and standard) means the suppliers don’t have to radically customize every item they produce. “We’ve got a great 0.5 U star tracker” is something they can advertise, not “We’ll design a great star tracker for you from the ground up.”
Sure. I just think it’s symbolic of the deeply entrenched culture — down in the very language used — that prevents iterative development.
COTS could be central to an iterative technology/capability development program. In fact it would be a perfect framework for such development. But I was pointing out that in practice, it wasn’t. Once the basic capability was demonstrated, it became a fixed supply program. The two programs, cargo and crew, were entirely independent, not cumulative.
Even their most innovative contracting system still shows the weird NASA bias against ongoing development.
Re: Falcon block 1 vs block 5.
But it’s worth noting that SpaceX has had to train its contract managers at NASA and DoD to understand and accept that process. The default position is to lock the design down at the beginning of the contract. The agencies had to change how they assessed programs to accommodate SpaceX’s iterative improvements. The assumptions baked into the very methods used to do contracting, even intentionally flexible commercial contracting, get in the way of improvements.
Would you agree that the early capsules – particularly Gemini – were improved a result of flight experience?
I think I would, although Apollo might be closer to what I mean by an evolutionary development. The vehicles used on J missions were improved from original designs, but (more to my point) the improvements and changes were not defined before they had experience from previous Apollo flights. I don’t think the Gemini hardware changed all that much from start to end. At least not based on experience from past flights. What they expected it to do, and what it did, certainly changed.
But the whole flow down from mission goals to requirements to designs and hardware is the current practice. In the 1960s, NASA was much more flexible and didn’t have the same processes.
Since reuse of the 2nd stage isn’t likely to be completed by the end of the year, does this mean that Falcon 9 won’t support 2nd stage re-use? Will that be for Falcon 10?
Second stage recovery would be awfully difficult. It is, by definition, going at orbital velocity before it can turn back. Propulsive reentry would require a huge amount of fuel, and non-propulsive recovery just isn’t an option. I really doubt it would be worth it. There could be reuse options, by leaving it on orbit. There were studies of leaving Space Shuttle external tanks on orbit, either to build a tank farm, as pressurized volume (a reboot of the old, Skylab wet workshop concept) or simply because any mass on orbit could be useful. But I think all that requires quite a bit more orbital infrastructure than we currently have.
For future launch vehicles, I’m fairly sure there won’t be a Falcon 9. I’m not sure where the “Falcon” part of the name comes from, but the “9” definitely refers to the number of Merlin engines the first stage uses. For a company (or CEO) who came up with names like “Falcon”, “Merlin” and “Dragon” (after Puff), I sure they will come up with a fine name for their next generation launch vehicle.
Falcon refers to the Star Wars “Millennium Falcon” spaceship.
> non-propulsive recovery just isn’t an option.
The space shuttle is an existence proof already that one can build and fly a reusable stage to and from orbit.
Now, you probably mean something like “all the recovery hardware built in to a second stage that would work on top of a Falcon 9 and have significant payload — is very hard to imagine.”
What I meant was that the Falcon 9 upper stage can’t be modified for non-propulsive recovery. You could certainly design a whole new vehicle which could reenter without thrust, and put it on top of a Falcon 9 upper state. But that’s essentially what you’d have to do. Less major modifications to the existing upper state aren’t going to do it.
Yeah, I agree. Major mods required.
A hypothetically reusable Falcon 9 upper stage would consume a very large percentage of the available payload-to-orbit. The small payload remaining would probably not serve the market very well.
It might be more interesting considering the second stage on FH. The remaining payload there might be enough to close the business case for a reusable second stage for small LEO sats.
I agree it is difficult. I think one of the major “inventions” by SpaceX is using rocket exhaust as a re-entry shield. NASA appears to be looking at it for Mars landings. It would be interesting to see how feasible it would be for earth re-entries. If launches are much cheaper, it may be worth having a larger upper stage with the fuel for a propulsive re-entry.
NASA investigated (at least on paper) using rocket exhaust as a heat shield years ago, but it was not particularly effective. It’s my impression the primary strategy with the Falcon is to use thrust to decelerate the stage prior to entry so that entry velocity and thus thermal loads are reduced.
Upper stage reuse might come with Falcon Heavy since for “lighter” satellites the second stage might have enough mass and fuel reserves to attempt recovery. If not, there are plans for a LOX/methane upper stage based on Raptor technology (but smaller than a full sized Raptor). That stage should have even better performance than the current LOX/kerosene upper stage. More performance means even higher mass and fuel reserves for “lighter” satellites.
Like most governmental agencies, NASA is trying to get funding (government contracts) for their buddies in Florida, Alabama, Mississippi, Texas and California and trying to guarantee results. It seems that the private sector has a better handle on this. Create systems, lay out achievable goals and then get to work.
Shuttle, too, had a slight glimmer of the idea of it but the few accommodations & upgrades that took place were so ‘governmently handled’ in most instances that they never made a monster impact on operational overhead, etc.
Tiny things, like the incorporation of direct payload interfaces via the portable computers (vs the standard payload s/w in the General Purpose Computer mainframes), were in most dimensions responsive. Don’t forget the losing of the white paint on the ET to up payload a fraction (and save some cost), etc.
Something to consider, I think, should be the revisions of or additions of entire missions to accommodate initially unexpected circumstances (including previous lack of success on earlier missions; TSS & the satellite repair & recovery flights come to mind). Not hardware upgrades like you are describing, but to some degree emblematic of system responsiveness.
There were some others, but…certainly not enough and rarely handled with efficiency. The worst page of any book about the shuttle is near-the-last-chapter or appendix section with the panalopy of major system upgrades like the LRBs, Shuttle-C, Orbiter-II, the ‘passenger module’ for the payload bay (a la Moonraker, minus the Marines)… that never came to be.
Sigh.
Hindsight is always 20/20 but imagine where we might be if the Shuttle was evolved rather than scrapped. Would the notional designs for a fly-back first stage ever become a reality? Hmmmm, sounds familiar.
That would have required starting with a “de-evolved” shuttle.
Ie, the STS should be seen as the final version. So the question would be, would that final version have been better/cheaper/more-capable/more-flexible/more-reliable if it had been developed over several versions rather than leapt to in a single step.
Once you had the STS, there was little room to evolve. It was too late. Not only was everything was balanced against each other, like a house-of-cards; to change any one thing meant changing everything. But also the operation costs were so great, there was no budget for a major upgrade.
Launching a 100 ton cargo container to put 25 tons of cargo into space…was ….not smart..especially when the decision was made to build it…
To be fair, it could have made sense. If the net savings from reusing the three SSMEs and the avionics exceeded the cost of launching 75 tons to low Earth orbit, then it would been a smart idea.
But, then, why be fair? The estimated savings from reusing the orbiters, the cycle time, and basically the whole business plan, were delusional in hindsight and probably should have been greeted with extreme skepticism at the time.
I don’t think our material sciences were quite there to do a space shuttle. A clean sheet design today .. launching on a FH… what a difference in costs…
Well, would you settle for an X-37B launching on a Falcon 9? Then maybe using flight experience to evolve it into a larger vehicle launching on a Falcon Heavy?
Of course, the X-37B is military, so cycling the flight experience into a new design might not be easy.
Fact is, what Space X is doing now is what space is and should be all about-developing the industry and the marketplace.
Relying on a Federal bureaucracy to manage a space transportation business is NOT the way to advance industry or make money. NASA probably has a role, but designing, building, managing and operating spaceships is NOT IT.
SpaceX is in the midst of a grand experiment, one that I dearly hope will be successful, but only time will tell. They have to balance many elements, from deeply technical to public confidence. So far they have done spectacularly well and I hope it continues, but the road is long and there are many treacherous areas ahead.
It would be wise to remember that the reason traditional launch vehicles have been so expensive, staid and uninteresting is the need for extraordinary reliability. Rockets use highly reactive propellants and operate at such high stress levels that relatively minor deviations from the norm often cause catastrophic failures. As a result, the best launch systems ever built have a failure rate measured in percent. Not one in 1000 or 10,000 as many would wish, but 1 in 20 to 1 in 50. This includes the Falcon 9.
The traditional approach to maximizing reliability has been to develop a configuration that works well enough and then perfect processes that produce exactly the same result time after time. This is boring and expensive, and really difficult when you have a field of vendors providing the components when small differences in manufacturing or materials can lead to failures. Space X has been bold in tackling these issues in many ways, including bringing so much of their manufacturing in house.
I look forward to the day when Space X has 100 flights in a row without a failure. That will show that they can achieve the needed perfection in the face of both physics and business pressures, including competition. When that happens they should open a university for reliable systems design because they will have changed the industry completely.
But then, they have already changed the industry even if they do not meet that reliability goal. They have introduced so many innovations that there are whole new degrees of freedom available to make a launch system cost effective. No doubt Musk will play with those degrees of freedom and demonstrate what is possible.
As for NASA’s approach, it is the result of many factors including politics. Many times I have discussed innovative ideas with NASA leaders and found they were blocked from even proposing things that would upset the Congress. Until they are given the freedom to pursue technical and programmatic optimization, the agency will flounder.
In the end, the space business requires people with a clear vision unimpeded by politics, ego, and fear. It is difficult and dangerous work (when humans are aboard), and anything that diverts attention from the perfection it requires hurts. There is no room for arrogance or timidity, only humility and measured confidence.
I appreciate Keith’s comments about NASA, but very sad. Perhaps they and others will learn a thing or two from Spacex. In their defense, doesn’t SLS stand for Senate Launch System and wasn’t it almost completely specced out by legislation? I think we can salvage this by replacing SLS with a Spacex contract and letting NASA do its Mars thing on purchased hardware. NASA still serves an important exploration purpose of doing missions with no expectation of commercial gain and only expectation of expanding mankind’s knowledge horizon.
When it comes to Blue Origin, I am from Missouri. Spacex had many, many early failures with Falcon 1 and Blue Origin still has a lot of failures and disasters ahead of it. Achieving the same success is not a given, and in my opinion, the odds are against them just as they were against Spacex. Just because Spacex beat the odds, does not mean Blue Origin will as well. I do not believe they have even launched to orbit yet. My old Estes rocket was reusable too.