The question is already being asked on social media: Was today’s launch a success? The short answer is YES! The longer answer is the subject of today’s post. Before I dive into that, some quick background given that the internet (particularly social media) is full of bots.
I covered aviation, space, science, and related topics for several years. Under Dave Dooling’s administration, I served as Correspondent-at-Large for SpaceWorld (article about as it is long gone) magazine for a few years, and worked a bit in radio. Hit the Readers’s Guide to Periodic Literature (hope it’s still around) for the byline C. Blake Powers. I later worked at the USAF Arnold Engineering Development Center (now Complex) and was there for the J5 incident. Said complex was/is the Free World’s most comprehensive testing site and could test at simulated altitude in a variety of the test chambers. After getting my Master’s, I went to work for Essex Corporation, where we were a subcontractor providing support to Spacelab through TBE. Among other things helped write a number of mission brochures, reports, and got to do a LOT of neat things. Left for a while, got asked to go to work for a company called CST working as a contractor for the Space Product Development Program (Commercial Space) where I was Director of Outreach and Education. Even was on a panel with Elon as I’ve noted a time or two before. As also noted previously, my thesis was The Soviet Watchers: A Directory of Western Observers of Soviet Space Efforts and it should still be available at the UTK library. Also, earned a pilot’s license, got to go through altitude training (and ejection training) and certification, and a few other things. So, not just an anonymous internet rando and things can be checked out fairly easily.
So, today was a success. The people at SpaceX really weren’t joking when they said if it cleared the tower it was a success. Given the number of rockets and rocket systems over the years that have taken out the launch pad on their first test, it really is a good thing. The first time you stack it all together and light the candle, anything can happen.
They got it off the pad. They got it up to a pretty good altitude. Then it went south. Happens. In fact, it’s a good thing when it happens during testing.
SpaceX is doing what should have been done by many others: they test. They test to destruction. The Starships that exploded in ground testing? Good thing. Lots and lots of data. They were not failures, each one enabled the next to be improved. Certain agencies and many companies don’t want to test to that extent, as they are convinced the public sees such as a failure when it is not so. Yes, I know there are idiots that do feel that way, but they have no clue about reality as a general rule.
Today, they got more data on the assembled vehicle and how it performed both on the pad and in use. They got reams of data on fueling and related issues. They got reams of data on each engine and how they performed together. They got reams and reams of data on major and minor systems. Data on the micro and the macro that you really can’t get except in flight. Ground-based testing can test individual components or systems. You can’t test something that large except in flight.
So, you do all the ground testing you can. You make each part as good as you can with that data. Then you integrate and launch to test. You do so knowing you are probably going to lose some of the test vehicles. As I said before, it is a good thing.
Had today’s launch gone perfectly and everything worked exactly as predicted, I would have been amazed, delighted, and concerned. On something this complex, if you have a perfect flight on your first test flight, smart people tend to ask if you really got everything perfect; or, if you just got lucky and missed something that is going to bite you in the ass later as the odds change? The latter is the safe bet, by the way.
Just a guess, but it looks like they had several issues. Several of the engines failed early. The complex separation maneuver did not go to plan. Obviously the stage separation systems did not work to plan. We will learn more in the days ahead, as it takes time to go through the massive trove of data from a test like this.
And that’s the point. That’s what makes today a success. The data gathered today is worth the cost of losing five Starships. With that data, good analysis, and good engineering, you redesign, refine, retool, and relaunch. Then you take the data from that launch and do the dance over and over again. It is an iterative process and if you think they aren’t doing it Falcon and other things, I’ve got a bridge for sale, cheap. It is the smart way to do it, and Elon is a pretty smart guy who also hires a lot of smart people to work for him.
So, unlike this morning’s memory, no rye today as there was no scrub, just a good test. If it had gone perfectly, I do have some Sazerac standing by but while I wish they gotten just a bit further, I’m delighted they got as far as they did and even more delighted at the data they got. It will allow them to get further ahead faster, and we need to head for the stars.
UPDATE: Go read this excellent guest post by Thomas Kendall over at Sarah’s place.
Getting hit by lightning is not fun! If you would like to help me in my recovery efforts, which include moving, feel free to hit the fundraiser at A New Life on GiveSendGo, use the options in the Tip Jar in the upper right, or drop me a line to discuss other methods. It is thanks to your gifts and prayers that I am still going. Thank you.
There’s a drone photo online already of the hole that the first stage excavated under the OLM, plus a wild video from a dashcam on a vehicle parked along the road fairly close to the pad showing lots and lots of flying chunks of concrete, which likely explains both the engine outs and any gimbal hydraulic issues that contributed to the progressive control divergence. The sustained roll visible early on once they went to the booster cam was an indication of control issues. Then at the point they were looking for MECO and staging it looked like either they didn’t get a full MECO shutdown or the computer was still trying to add some energy. One rumor is that the vehicle was still seeing too much atmospheric pressure to let it command separation, even if the body rates would have been in range to try.
All of which is a long way of saying the same as above: Successful Initial Flight Test, Lots of Great Data.
Might want a flame trench after all, though.
Heh. Think you’re right on that, think a pad redesign and rebuild is in order. :).
Your comment on something not going wrong is interesting. The first test of the complete Apollo launch stack, Apollo 4, was a complete success.
Why, yes, there are exceptions to every rule.
True. However, Apollo 6 did have engine failures that would have scrubbed a crewed mission. But as the test happened the same day as MLK’s assassination it was pushed out of the news cycle and largely forgotten.
The second test of the SV launch vehicle, Apollo 6, had extreme pogoing.
Latest armchair internet failure analysis is that the launch video shows both hydraulic power packs exploding at separate points during ascent, which would fully explain the loss of control. Apparently this is the last booster with hydraulic gimbaling, as they’ve already changed to electric in the next booster coming off the line, so next flight they get to look for different failure modes.
The real question is whether all that flying concrete debris is the root cause of everything that failed.
That really does look to be a major factor. Hope they share some of the findings, and any papers that come from this.
I always told young engineers that you know that Murphy really hates you when the first one works.
I LOVE that phrase!
NASA and DoD have become so rusk intolerant that a single flight failure puts a program at risk of cancelation. I see cases where the failure of a secondary component on a brand new system results in the government customer stopping all other work until root cause is identified and corrected. Program competitions are often won or lost when one company suffers a test failure. “Company A’s system works, and yours doesn’t,” after one data point. It’s great if you’re the winner, but supremely frustrating when you’re the laser. SpaceX is eating the rest of the launch industry for lunch because they are willing to test to failure. The legacy companies won’t risk losing a launch test vehicle.
Excellent points! NASA is so risk averse it’s amazing we have anything by them in orbit. I worked TSS-1 and TSS-2. Failure of TSS-1 was the result of a “safety” change with no mechanism test after. I still suspect it took blackmail to get them to fly the second mission given the reaction to the first failure. After the second failure, well, it won’t be NASA doing anything with tethers.
Musk observed after several earlier (explosive) Starship prototype failures that none of what turned out to be the causes of those failures had appeared on their preceding expected trouble lists.
I watched the launch, and the intentional destruction of the vehicle. I assume that SpaceX knows what it’s doing, but some things don’t make sense.
The Super Heavy first stage lost 5 engines shortly after takeoff… you can see from the SpaceX video that several engines have shut down. This obviously led to an asymetrical thrust condition that sent the rocket spiraling out of control. This engine is the same engine that been used previously… just more of them in this first stage… but 5 failures out of 32? 15% failure? Some failures are unavoidable, but this seems a little excessive with proven engines.
Are there provisions for the Starship to detach from the Super Heavy first stage if something like this goes wrong on a crewed takeoff? One would think that the ability to throttle back the engines, detach from the first stage and separate, and then fly the Starship back to a safe landing would be a necessary capability. Starship is supposed to have the autonomous ability to land itself… was this not enabled? We’ve seen how autonomous landings are hard and how many earlier SpaceX boosters were lost perfecting this capability. Why wouldn’t you want to test this ASAP?
I know, the Starship was announced to be a write-off before launch… but that seems to be a rather expensive and wasteful way to iterate towards success.
I know, that seems to be a rather expensive and wasteful way to iterate towards success… but the Starship was announced to be a write-off before launch.
FIFY
Think of all the expense and waste of enabling those functionalities in a test where they were not needed! Sure, throw in an x-million-dollar self-lander on a ship that will never land. Did they even test the life support systems? On a ship a crew would never ride? There is the question of time to be considered. They’re in a hurry and only have X man-hours to do it with.
This apparently is how they operate. If you or I knew how to do this, we would be outcompeting NASA and the ROW too. One needn’t worship at the temple of a god if he has feet of clay, I personally cannot stand those wisps of beard he sometimes affects, or else he needs to learn how to shave right, but if his rockets fraudulently or incompetently don’t work, then obviously no one is more harmed than he is.
On the bright side, it seemed that Starship , the upper stage, never tore up the pad like this, so at least the lander may be able to operate without a prepared surface. On the downside, and they absolutely needed to do it exactly as they did to learn the downside, because they needed to find whether an “un”prepared (because how are you going to prepare the surface of the moon or Mars) launch surface would affect the working of the booster, Super Heavy:
And a semi-prepared, concrete pad did get torn up, and so I guess they’ll either have to fix something, redesign it for softer T/Ls, or redesign the pad for harder T/Ls. Now they know.
Note that if it was a hydraulic failure like a chunk of concrete tearing a hose, such a complication will become impossible in subsequent versions, which will not be hydraulic but electric. Best part is no part.
I watched the launch, and the intentional destruction of the vehicle. I assume that SpaceX knows what it’s doing, but some things don’t make sense.
The Super Heavy first stage lost 5 engines shortly after takeoff… you can see from the SpaceX video that several engines have shut down. This obviously led to an asymetrical thrust condition that sent the rocket spiraling out of control. This engine is the same engine that been used previously… just more of them in this first stage… but 5 failures out of 32? 15% failure? Some failures are unavoidable, but this seems a little excessive with proven engines.
Are there provisions for the Starship to detach from the Super Heavy first stage if something like this goes wrong on a crewed takeoff? One would think that the ability to throttle back the engines, detach from the first stage and separate, and then fly the Starship back to a safe landing would be a necessary capability. Starship is supposed to have the autonomous ability to land itself… was this not enabled? We’ve seen how autonomous landings are hard and how many earlier SpaceX boosters were lost perfecting this capability. Why wouldn’t you want to test this ASAP?
I know, the Starship was announced to be a write-off before launch… but that seems to be a rather expensive and wasteful way to iterate towards success.
I think the engines were a result of the launch pad spalling concrete, to little suppression – no flame trench.
Pogo oscillation would cause that failure.
See what I wrote below – you either shut down the engine, or the rocket tears itself apart. But you shut down six engines…now you’ve got asymmetrical trust issues.
I know, that seems to be a rather expensive and wasteful way to iterate towards success… but the Starship was announced to be a write-off before launch.
FIFY
Think of all the expense and waste of enabling those functionalities in a test where they were not needed! Sure, throw in an x-million-dollar self-lander on a ship that will never land. Did they even test the life support systems? On a ship a crew would never ride? There is the question of time to be considered. They’re in a hurry and only have X man-hours to do it with.
This apparently is how they operate. If you or I knew how to do this, we would be outcompeting NASA and the ROW too. One needn’t worship at the temple of a god if he has feet of clay, I personally cannot stand those wisps of beard he sometimes affects, or else he needs to learn how to shave right, but if his rockets fraudulently or incompetently don’t work, then obviously no one is more harmed than he is.
On the bright side, it seemed that Starship , the upper stage, never tore up the pad like this, so at least the lander may be able to operate without a prepared surface. On the downside, and they absolutely needed to do it exactly as they did to learn the downside, because they needed to find whether an “un”prepared (because how are you going to prepare the surface of the moon or Mars) launch surface would affect the working of the booster, Super Heavy:
And a semi-prepared, concrete pad did get torn up, and so I guess they’ll either have to fix something, redesign it for softer T/Ls, or redesign the pad for harder T/Ls. Now they know.
Note that if it was a hydraulic failure like a chunk of concrete tearing a hose, such a complication will become impossible in subsequent versions, which will not be hydraulic but electric. Best part is no part.
Update: You seem to have a sock puppet named Obijohnkenobe who posted the identical comment. There would be no point in replying twice. Goodbye.
Update: Now I just posted the same comment twice. But I’m the same person so a mistake, not sockpuppeting.
Isn’t it odd that there seems to be a motivation by some people to maximize fear, uncertainty, and doubt about SpaceX and Starship, the nation’s one hope for orbital salvation and beyond. That’s not worship of Elon Musk. That’s cold fact that nobody else is even close.
If not him then China. Pick.
Your comment about getting it right the first time makes me even more amazed that STS-1’s liftoff and assent were successful.
That’s a double-edged sword, though: the Internet Geniuses will point to STS-1 as an example of NASA doing it “right”, and SpaceX being an extravagant waste of money – despite the documented $93 billion cost overrun, and it’s 6 year delay from original launch date in 2016.
Plus, as it was noted, having nothing go wrong on the first launch of a new system just means that Murphy is quietly laughing in the background.
STS-1 had the blowby-possible o-ring design, as well as fundamentally the same foam (albeit even heavier with the white paint) as eventually caused loss-of-crew level failures.
Would they have launched STS-1 on as cold a day? No way, given flight test folk were in charge of that first flight. And likely that foam was way more inspected being it was the first flight, no unmanned test being possible. But the faults that eventually killed astronauts were on that stack – they just failed to fail that day.
There’s also a story about a chunk of pad metal gone after launch. It was found lodged in the nozzle of a SRB found in the ocean when they tried to insert the plug to recover them. Solution was to plus up the water wash significantly before the second launch. If that sucker had bounced in the wrong direction and penetrated the lower ET dome, really bad things would have happened.
For the record, I never found anything in the public other than war stories about this one. Cheers –
How much wasted qua wasted, anyway? How much does one ship cost to build?
The rocket stayed too long on the pad after the engines began to fire, FOD was undoubtedly a major problem. And speaking ofthings that need work — the “Flight Termination System” needs to be redone, right after the Range Safety Officer is fired for not terminating the flight earlier… say at the end of the first loop, maybe even the second. I’m pretty sure the rocket was terminated in the third loop…
Flight termination for rockets launched from US federal ranges is converting over to autonomous systems. Generally there is a parallel computer system that tracks the flight with dedicated sensors and fires the destruct charges if it detects the rocket going outside a pre-programmed flight corridor. For a first experimental flight the corridor will generally be programmed wider than for a mature product – essentially, trading regulatory inconvenience for data collection.
I’ve worked on several D&D programs during my career. One of the first principles of testing is:
“TEST FOR STRESS, Not Success.”
I could always write a test that would seem reasonable and pass. But you learned nothing.
Any good engineer is known for breaking things.
Good post, but I have a couple of nagging concerns.
I would have tested super heavy and starship separately before integration. Iron out the bugs separately. This is a first launch for super heavy.
Secondly, how it failed is disconcerting. There are a million advantages to using many smaller engines versus fewer big engines, but one huge disadvantage – pogo oscillation. Small increases in engine burn increases back pressure against the fuel coming into the engine, reducing engine pressure, causing more fuel to come in and increasing engine pressure again. The engine either shuts down, or the whole rocket shakes to pieces or starts tumbling. This is what wrecked the N1 Russian moon rocket, and why Nasa is so in love with solid rocket boosters (which avoid this problem).
The starship booster has a record number of engines – 33. Pogo oscillation increases with increasing the number of engines, and gets harder to control. Imagine each engine, individually, having issues contributing to the whole. A small problem a rocket with 3-6 engines might have become unmanageable with 33 rockets.
It appears the failure might have been caused by exactly this problem. We saw 6 engines fail during flight, and the rocket tumbling before destruction. That is very much a pogo oscillation symptom.
Maybe that get’s ironed out. But…maybe not. A LOT of rockets have been destroyed by this problem, and a lot of rocket engineers have said it is unsolvable. I’m not that negative – but here we are, first launch, and the problem is the same one we have seen since the 1950s…before even designing a rocket with this many engines, I would expect this would be the first problem you solve.
This launch did eat the pad. Find a picture of massive hole the booster made under the stand. A simple water deluge (probably?) isn’t going to solve that problem.
If the current grapevine rumor turns out to be true that none of the engine-outs were debris-related, you’d have to conclude the concrete-and-sand exhaust energy dissipation system worked great!