SpaceX’s Starship V3—still a work in progress—mostly successful on first flight
- Thread starter JournalBot
- Start date
Perhaps you are right.
However, that's not the way I'd bet.
1) Raptors generally appear to work. They have very good Isp, excellent thrust/weight ratio, and are working their way up in performance, down in weight, and (presumably) down in cost. Overall, SpaceX has made steady progress over the last several years.
2) The heat shield appears to work adequately for orbital missions. Every mission that has reached orbital-energy-in-a-non-orbit that maintained RCS control has re-entered successfully and maintained control until (pseudo-)touchdown. This isn't proof it can or will work well for direct re-entry, but it's as good as, and in many ways better than, any other reusable orbiter has demonstrated.
3) The new launch pad appears to perform very well, and by all appearances looks likely to ultimately be able to perform multiple launches with trivial or no refurbishment.
4) Landing catches for the booster have been demonstrated. No failures... so far.
5) Propellant transfer experiments have been successfully performed during test flights. (Details unknown.)
6) Despite individual engine failures, booster and Starship flight control systems have maintained control of the ascent and descent as long as some engines and the RCS system have been able to provide control.
These, to me, indicate that the major problems preventing the creation of a fully reusable two-stage-to-orbit rocket all have practicable solutions. It's possible that SpaceX may not be able to assemble the components into an economical, working rocket. But, given SpaceX's demonstrated prowess at solving difficult problems, this seems unlikely.
That SpaceX is only now, with Version 3, getting enough infrastructure in place to start attempting bulk propellant transfer on orbit, indicates that they're not terribly far out ahead of their skis. They also now have a "pez dispenser" that can quickly dispatch flat-pack satellites on orbit, and their test satellites, as launched with IFT-12, strongly suggests they're likely nearing readiness with their next-generation Starlink satellites.
I can't speak to Musk's financial shenanigans with SpaceX. To me, that likely constitutes the largest threat to the continued development of Starship into a major step forwards in human spaceflight technology. Such a failure would IMO be a major disappointment. But I also believe that others now know or believe that the problems SS/SH were intended to solve have genuine solutions, and that someone, possibly a government (or two...), will spend the money to create such a system. That there are now experienced aerospace engineers who could, with suitable funding, create such a system appears inescapable.
Finally, as a retired engineer, it is not at all unusual for large, complex engineering projects to have multitudes of bugs and failures during development. It also is not unusual for them to take longer than expected to be delivered. And the bugs and failures are often not at all indicative of final product behavior. This is true also of non-rocket engineering products and projects. For Starship/Superheavy, there isn't any evidence I'm aware of that points to possible showstoppers: all the fundamental issues appear to have demonstrated, workable solutions.
Upvote
25
(25
/
0)
That will happen eventually. The question is when. Specifically, whether it will happen as soon as Artemis IV. I think the suggestion was that Starship won't be mature enough, the design stable enough. If refurbishment takes 6 weeks, then a weekly launch cadence would need 6 tankers for, say, 14 launches over 14 weeks. That's a fair amount of specialised vehicles to produce, update, and store. If they want to launch twice a week to halve the time for boil-off in the depot, then they'd need twice as many tankers in the refurbishment pipeline.
Another way of looking at it: If a dedicated tanker is 11% more efficient at delivering propellant, you can use 9 tanker launches to deliver the same propellant as 10 cargo Starship launches. If each launch costs $20M, and each tanker is $30M, you need around 15 tanker launches to break even. If each mission reuses tankers 3 times, that's 5 missions. At one mission per year, 5 years. Maybe these numbers are way out, but it's not looking compelling.
Upvote
3
(3
/
0)
There were a couple of failures on flights that were supposed to be catch, but they aborted before being close to the tower.
Upvote
5
(5
/
0)
The mars colonial transport was a much larger design made of aerospace aluminum and carbon fiber. The BFR was the same. Development consisted of raptor development and building an enormous carbon fiber tank that was junked. All the while Falcon 9 and falcon Heavy development consumed almost all resources.
Starship, an entirely new all stainless steel design was started at end of 2018 after Falcon Heavy entered service. It was the first credible design for launch through re-entry and the first to get significant R&D investment.
And super heavy did not fail because of an engine failure, it failed due to propellant delivery failure likely due to an all new propellant delivery system. As usual with the first two versions, SpaceX will likely have this fixed quickly.
Upvote
11
(11
/
0)
I think I am getting an extreme case whiplash on what exactly constitutes success or failure these days.
Upvote
1
(3
/
-2)
It's basically always this way with test flights (as opposed to operational flights). This flight certainly missed some key objectives (demonstrating the 33 engines boost back and precise booster landing on water, demonstrating in-orbit relight of a Raptor), but then the second stage landed successfully and the heat shield as well as the flaps looked really good this time.
IMHO this flight looked better due to successfully landing the second stage then it actually was. They will have to repeat the very same flight next time without being able to move on to the next objectives with it as planned (like going to full orbit or trying a booster catch). Which again pushed everything else (catching the ship, deploying actual payloads) one flight to the right.
So in isolation it was a success if you just look at the end result (landing the intact ship successfully on the water). But as a part of a test campaign it was more of a failure since they now have to repeat it instead of moving on to testing other things that depend on that.
Upvote
7
(9
/
-2)
Sure, that is both a nuanced and reasonable way to look at it. Wasn't really what I was commenting about though
Upvote
-3
(0
/
-3)
Sure, that is both a nuanced and reasonable way to look at it. Wasn't really what I was commenting about though
Well, if you're asking (I'm guessing here now) why SpaceX celebrated this as an success: This flight was very close to end up as an utter and complete failure. When after the hot staging not only the booster did not start the boost back burn and all engines flickered out, but 35 seconds later also the ship lost an engine you could basically hear all hearts sinking there. This looked too much like they were about to lose both stages, which would have been a very bad start for v3.
I certainly was close to just closing my browser tab at this moment. When the ship soldiered on and finally got through and landed this felt like a success. Things can be both a success and a failure in different ways.
But probably this also isn't what you were commenting about.
Upvote
-7
(1
/
-8)
I wasn't asking anything. I was simply commenting about all the folks constantly moving the goal posts around in these threads about what it means for various things to be a success or a failure.
Like saying not making orbit is a failure when the goal has been to test reentry. People seem to think they get to decide the success/failure mission parameters instead of SpaceX. The number of absolute statements made around here based on personal opinion without actual knowledge is just kind of absurd sometimes.
Upvote
12
(13
/
-1)
This is a tough crowd. This was an orbiter design that has never flown before, atop a booster design that has never flown before, atop 39 redesigned engines that have never been flown before, and it was very close to a complete mission success. I was hoping for complete success, but not expecting it, and thus was pleasantly surprised at the outcome.
Frankly I"m a bit peeved at the FAA for declaring a mishap when this was a completely predictable and foreseen possible outcome. I don't understand why a company like SpaceX can't declare an experimental flight plan that says, "Plan for outcome A, B, or C" and then be able to proceed as long as one of those 3 outcomes occurs.
Frankly I"m a bit peeved at the FAA for declaring a mishap when this was a completely predictable and foreseen possible outcome. I don't understand why a company like SpaceX can't declare an experimental flight plan that says, "Plan for outcome A, B, or C" and then be able to proceed as long as one of those 3 outcomes occurs.
Upvote
12
(12
/
0)
They do have the ability to do that to an extent. Like they can say we will try to land on the tower if everything looks good, otherwise we land at this spot in the water.
However trying to fire up your engines for boost back, having them fail, taking out other engines, and then not hitting the water where you said you were aiming for (even if still in the exclusion zone), is usually going to be a stretch for that.
The mishap investigations don't typically show them down. SpaceX would be doing one internally anyway, they're just sending the results to the FAA who will review and approve it.
Upvote
17
(17
/
0)
I'm sure I've been ninja'd here but, the fact that NASA has not flown crews on its own spacecraft since 2011 is not due to any technical failure on NASA's part, but due to a deliberate policy decision on the part of the United States government to stop launching its own rockets and hire that service done by commercial providers. NASA did not design and build the SLS, Boeing and its subcontractors did, within the very tight constraints laid on them by Congress. All SpaceX does is launch the rockets, NASA does the exploring.
Upvote
-2
(3
/
-5)
The two ways the IFT-12 test flight failed are (IMNSHO):
1. The flight did not test microgravity engine restart.
2. The flight did not test the booster landing flip and engine restart.
The former (1) was not tested for unknown reasons; most likely the second stage engine out event on ascent precluded the test.
The latter (2) was not tested, almost certainly due to the substantial engine failure(s) during the booster flip & boostback burn failure.
As you've pointed out, these will need to be retested in the future, setting SpaceX back, perhaps by as much as a launch. :-(
Conversely, however, pretty much everything else in the V3 changeset was tested:
I. V3 engines were tested in the high-vibration, high-noise, high-everything launch environment.
II. Launchpad two and its many ground support systems were extensively tested.
III. The new hot-staging and booster flip maneuver was likely too aggressive, and SpaceX now knows there's a problem there.
IV. There were individual engine outs in both the booster and Starship, and SpaceX will need to examine the telemetry to diagnose and correct the problem(s).
V. The integrated engine shielding may not be adequate for SpaceX's wants, but did prevent/avoid booster RUD; telemetry should provide information.
VI. The Starship flap changes appear entirely successful.
VII. The new propellant feed systems appear to work well (at least for launch).
VIII. The avionics changes (perhaps excepting any internal to the Raptor V3s) appear to be successful.
IX. The new three-gridfin system appears to provide adequate and likely improved booster control.
X. The integrated hot-staging ring appears to work.
XI. The new booster and Starship fueling ports and quick disconnects appear to work.
XII. The reworked Starship door worked.
XIII. The upgraded "PEZ" dispenser works, and the dummy and "dodger dog" satellites appeared to work well (and reportedly allowed other tests of the Starlink satellites and network).
XIV. The simplification/removal of the temporary on-board fire suppression system appears successful.
Upvote
11
(12
/
-1)
BO NG-4 just blew up on the pad. It’s not just SpaceX that blows up equipment, space is hard. And this was an in service rocket that BO lost, so its design should be far more stable.
Upvote
11
(11
/
0)
I believe they can and do. Unfortunately, this is one they didn't plan for. Still, I find it interesting that it took the FAA a few days to decide to call it a mishap.
Upvote
2
(2
/
0)
TLDR: I think Starship 12 went pretty well overall, but that doesn't mean you shouldn't do post-launch investigations into what went wrong.
"The FAA determined that the loss of the Super Heavy booster is covered by one of the approved test induced damage exceptions requested by SpaceX for certain flight events and system components," the agency explained. "The FAA evaluated each exception prior to launch approval and verified they met public safety requirements."
The issue is that SpaceX can't give an opened ended "well it might not work and we don't know how it might not work and we don't know what will happen" in that case, the mishap investigation is required to be done to figure out what happened (and for the results/conclusions of that investigation by SpaceX that to be reported to the FAA).
https://www.faa.gov/space/compliance_enforcement_mishap
It list that if any one triggers a legally required mishap investigation.
The FAA regulations describe nine events (see below) that would constitute a mishap (14 CFR 401.7). The occurrence of any of these events, singly or in any combination, during the scope of FAA-authorized commercial space activities constitutes a mishap and must be reported to the FAA (14 CFR 450.173(c)).
As @MLMichael said SpaceX can declare an flight plan with some known failure modes. That's exactly what SpaceX did for Starship 9 - because although they did a mishap investigation for the Starship failure, SpaceX was not required to do it for the Booster failure because they got a pre-approved exception.
"The FAA determined that the loss of the Super Heavy booster is covered by one of the approved test induced damage exceptions requested by SpaceX for certain flight events and system components," the agency explained. "The FAA evaluated each exception prior to launch approval and verified they met public safety requirements."
The issue is that SpaceX can't give an opened ended "well it might not work and we don't know how it might not work and we don't know what will happen" in that case, the mishap investigation is required to be done to figure out what happened (and for the results/conclusions of that investigation by SpaceX that to be reported to the FAA).
For better or worse, FAA is a government entity and I'm sure there was a lot of paperwork and red tape that needed to be done before formally announcing it. I think technically the FAA can decide the anomaly wasn't severe enough to trigger a mishap, but they have to actually do the post-launch data/telemetry analysis to list exactly what aspects of the launch SpaceX need to investigate and report back to the FAA. Here's the FAA page on Mishap
https://www.faa.gov/space/compliance_enforcement_mishap
It list that if any one triggers a legally required mishap investigation.
The FAA regulations describe nine events (see below) that would constitute a mishap (14 CFR 401.7). The occurrence of any of these events, singly or in any combination, during the scope of FAA-authorized commercial space activities constitutes a mishap and must be reported to the FAA (14 CFR 450.173(c)).
- Serious injury or fatality
- Malfunction of a safety-critical system - The failure of the Booster engines re-ignite (unless SpaceX explicitly says it might fail and ask for prior exception) is considered a safety-critical system as the envelope of descent with only grid-fins is much narrower.
- Failure of a safety organization, safety operations or safety procedures
- High risk of causing a serious or fatal injury to any space flight participant, crew, government astronaut, or member of the public
- Substantial damage to property not associated with the activity
- Unplanned substantial damage to property associated with the activity
- Unplanned permanent loss of the vehicle - SpaceX planned to lose the Booster, but only after it softly landed in the ocean)
- Impact of hazardous debris outside of defined areas
- Failure to complete a launch or reentry as planned - Both Starship and Booster had anomalies in launch and re-entry that were unplanned (or specifically considered as as failure mode).
Upvote
10
(10
/
0)
Eric Berger has a front-page article up on it. Lots of activity over there for the last few hours.
Upvote
4
(4
/
0)
Fair assessment, but I'd say that 1 out of 33 engines failing in the booster after 1:42 min seems to indicate that Raptor 3 may not be as reliable yet as it should. Even if we ignore the 1 of 6 engines in the ship also failing after 35 seconds (which may have been an effect of the somewhat botched hot staging) a failure rate of 3% is way too high for flight hardware and 33 engines isn't a really small sample. Before this flight I was really expecting with Raptor 3 to not see a Raptor failing anymore for a very long time. Engines failing has become all too normal during the test flights, to the extent that not a single engine failing (or not igniting for the boost back or landing) is even a remarkable exception.
But yes, it was a good test for Starship v3, most of the new stuff worked fine. I wouldn't just scoff though at the fact that everything will now move one flight to the right. This will ripple into the future and also delay the first operational flight and with this the opportunity to earn money with it instead of just spending it. Being able to launch their newest and most capable Starlink satellites only later will cause some non-trivial loss of revenue down the road.
Upvote
1
(1
/
0)
That's why these test flights exist: SpaceX, so far as I am aware, has no way to test individual Raptor engines in a simulated launch. They may have approximations, but those approximations aren't worth much when the launch vehicle is substantially new or different.
I was not scoffing; there are many things that may not move one flight to the right. Every test flight has and will test a variety of systems that are new, have changed, or may be affected by other systems that have changed. In traditional software development, every change made to a production product used to have something like a 15% chance or more of causing a new bug. I doubt hardware development is very much better.
The only true test failure is a test that fails to test the condition(s) it was intended to test.
Upvote
9
(9
/
0)
Yes, simulating launch conditions on the test stand is hard and in some ways impossible. Still, if 32 engines work fine during the launch and one fails you either have problems with quality management or the engines are working with such razor-thin margins that it's random chance which one fails.
They really have to work on this, because with the Artemis HLS these very engines will have to work not only for the second stage burn to LEO but then also for TLI, braking into lunar orbit, landing on the Moon and launching the crew back to lunar orbit. And hardware-rich development or not, they can't afford to try this a dozen times until it works. They have to not only test, but demonstrate near to 100% reliability of their engines or NASA will never risk to put a crew on HLS. It's crazy enough already in the best case.
Upvote
-6
(0
/
-6)
At this point after four attempts, their design should be pretty stable as it has proven feasible in all parts of the mission profile, albeit not all in one mission. That said, won't have real clarity on this until they have root cause, or a highly likely one.
Upvote
7
(7
/
0)
The other possibility when you have brand new engines and an engine out capability is that the engines have rather tight operating constraints on them until they gain more flight history. Having an engine cutout early is much preferable to a RUD. Anyway they went through their with raptor 2, and I see no reason they can't resolve it with raptor 3.
Upvote
10
(10
/
0)
Bill Swallow
Ars Centurion
I would have added:
XV. The largely successful real-life test of the Ship's heat shield; it appeared to show significant improvements in performance over the earlier versions.
Upvote
8
(8
/
0)
I wonder if their flight software has the ability to "take a risk" and suppress the most conservative shutdown conditions after enough engines have been lost that the next loss will trigger a LoM ... at some point, shutting down an engine that dooms the craft is no longer an entirely conservative decision; it's dooming the mission to avoid a chance of loss. Alternatively, if too many engines have shutdown, they could attempt restart on one or more of the failed engines; as long as >0 engines are thrusting the propellant will be settled in the task, and if all engines are offline, they could probably fire RCS for ullage.
The main counter-argument I can imagine is that it might be preferable to have an intact Starship enter and crash into the ocean rather than detonate and reenter as a wider cloud of debris that causes greater uncertainty for commercial aircraft.
A second potential counter-argument relates to spaceflight engineering culture, including at SpaceX (which has a unique perspective on risk that appears looser on the surface, especially during testing, but which resulted in F9 having the best operational track record of any launch vehicle ever): I just can't think of many examples where a spacecraft was programmed to take desperate gambles in service of saving the mission. Most space engineering organizations seem to prefer focusing their development effort on the bolstering the primary redundancies that prevent them from getting into a desperate condition in the first place. That, and every line of spaceflight code has to be path-complete tested, so engineering complex restart logic that almost never runs in practice is a non-trivial effort in service of questionable levels of risk reduction. And you quickly lose all the benefits if you accidentally trigger your desperation logic in a non-desperate scenario. Some notable exceptions to the apparent principle of "no desperation engineering" would include the Space Shuttle abort modes, several of which seemed quite ... optimistic (desperate). And SpaceX Dragon, after the
Last edited:
Upvote
3
(3
/
0)
SpaceX Dragon was altered to allow for recovery after launch failure subsequent to the CRS-7 failure. There's no LES on Cargo Dragon flights.
Another example of last-ditch recovery attempts is the current Crew Dragon - the Super Dracos can attempt a last-ditch retro-burn if the parachutes have completely failed.
Upvote
8
(8
/
0)
Yes, it's easy to miss it because nothing dramatic happened, but nothing dramatic happened during Starship reentry.
I'll never forget that epic moment on flight 4 when Starship landed with 2/3rds of the flap ablated away ... I was watching on the edge of my seat as the flap glowed and melted, waiting for the fireworks to start, sorry, "remaining vigilant to characterize the exact vehicle failure mode in response to traumatic control surface ablation", and I just lost it when all of a sudden the remaining 1/3rd of that flap articulated wildly to compensate for its reduced surface area and the craft stuck the landing. I was staring in disbelief not quite understanding why the telemetry was claiming zero altitude but the video was still rolling, when I noticed that there were wave crests lapping against the bottom of the vehicle through the narrow spot of visibility through the molten metal sputter on the camera lens.
Upvote
13
(13
/
0)
Thanks, fixed. Ah, arming the Super Dracos on parachute failure is smart.
Upvote
1
(1
/
0)
Thomas Cochrane
Ars Centurion
That's what prompted my original question. I didn't think they were in a position for a rapid turnaround because they've not practiced it and have been more concerned with the safety and status of the returned booster, thus far. If Starship can safely loiter in orbit, it's clearly not a concerning issue.
Upvote
1
(1
/
0)
I think it has been covered, but almost certainly Starship returns are likely to be after ~12 or 24 hours or some multiple of them, and also depending somewhat on the orbital inclination. Basically, there should be one or two times per day when the orbital track flies over any given point on Earth that it flies over at all (not generally 12 hours apart for locations off the equator). Or, perhaps more visually, when the orbital plane, which is mostly fixed in space, has any given Earthly location rotate through it.
Upvote
7
(7
/
0)
I was speaking of HLS landing on the Moon. You want to have them try this a dozen times with a dozen tanker flights each until they get it working?
With or without crews?
Upvote
-1
(0
/
-1)
The timing will be more important when they start using a Depot. I expect they'll choose an altitude such that they get an integral number of orbits in a sidereal day, so that the Depot will be passing overhead every other time the launch site passes through the orbit plane. That's when the next tanker is launched, and shortly before the previous tanker lands. I suppose they could launch two tankers on each pass, and recover two, if they have enough launch pads, landing pads, and confidence. I'd be surprised if they can arrange for the Depot to also pass over a second site at one sidereal day intervals.
Upvote
6
(6
/
0)
I imagine Starship will be programmed to use vacuum Raptors for the flip and landing burn if all the sea-level engines fail. As I understand it, there's a chance vacuum engines would survive flow detachment even if you wouldn't want to use the them again after. This might only be for valuable cargo, like crew, though. Otherwise it might be better to just write off the Starship by having it crash into the ocean rather than risk letting it near the pad.
Upvote
2
(2
/
0)
A week in retrospect it is an unexpected - new rocket - and relative success - in relation to the Begin Over (BO) explosion. But now the pressure will increase on SpaceX (and, to be fair, on BO).
And now they have announced that they aim to participate in M&M (Moon and Mars) flybys. https://eu.usatoday.com/story/news/...urrency-billionaire-mission-mars/90292894007/
? SpaceX allows others than NASA to do the exploration:
https://en.wikipedia.org/wiki/Polaris_Dawn
https://en.wikipedia.org/wiki/Fram2
And now they have announced that they aim to participate in M&M (Moon and Mars) flybys. https://eu.usatoday.com/story/news/...urrency-billionaire-mission-mars/90292894007/
Upvote
1
(1
/
0)
No, that's not going to happen. RVacs can't gimbal. At low speeds during landing, gimballing the engines is the only way to maintain control over the Ship (aerodynamic surfaces become ineffectual at low speeds, and RCS thrusters are far too weak to make any difference.)
Upvote
3
(3
/
0)
BO watched Ship’s tiny little mushroom cloud after its RSD and said, “Hold my Beer.”
Upvote
-4
(1
/
-5)
I'm not at all sure of this.
Originally, vacuum Raptor engines were, IIRC, not going to be throttleable: full power or nothing. However, that seems to have been relaxed; sources I find say that both Rvac and sea-level Raptor engines are throttleable through their whole range (40-100% thrust). While Raptor engines probably aren't as throttle-responsive as Merlins with their pintle valves in the combustion chamber, it may well be possible to use throttle and multiple Raptors to flip Starship. However, it almost certainly will not be possible to land it that way due to the minimum thrust requirements: it would have to be a "suicide burn" at an altitude where there isn't enough downward velocity to reach the ground. (And would likely fail anyway...?)
It's undoubtedly academic.
Upvote
0
(0
/
0)
An F9 uses a "suicide burn" and must hit zero altitude and zero altitude rate at the same time. Starship Rvacs are not centered. They would either have to finish with two Rvacs (a suicide burn with even higher acceleration), or with only one Rvac (and hit zero altitude, zero altitude rate, vertical orientation, and zero rotation rate all simultaneously). Not sure it would have the command authority for that. And they certainly wouldn't try unless there was a crew aboard.
Upvote
0
(1
/
-1)