So, I’m in Titusville, FL for a conference this week, and the news just reported that there’s a launch planned for today! If I understand right, this is the first launch of a 3D printed rocket, and simultaneously the first launch using a LOX/CH4 engine!
I think liftoff is around 1300 EST, so if we don’t break up the discussion, I’ll politely excuse myself to go check it out. I wish I could take video for y’all, but my company phone had the camera disabled.
It’ll be cool to see in person, though!
Tripler
T-minus 4 hours of PowerPoint boredom. . .
Sorry to hear you’re in Florida, cool that you’re possibly going to see a launch, Bayes says even odds it is destructed while recent history with new commercial launchers is more like four out of five, but fireworks are fun, too, if you aren’t downrange of them.
Has anyone, ever, had successful first launch of a first vehicle? I’m excluding first launches of second+ vehicles (Falcon 9), vehicles inherited from other companies (Atlas V), or vehicles developed by companies assembled from other companies with flight experience (Vulcan when it actually flies).
Yeah, it got scrubbed today. I haven’t heard it its been rescheduled for tomorrow, but apparently there is another launch tomorrow. I think I heard tomorrow’s launch is a replacement crew for the ISS.
The ISS crew launch (Crew-6) was last week. But there is one tomorrow for a batch of OneWeb satellites. There’s also a cargo mission to the ISS on the 14th.
OneWeb is a competitor to SpaceX’s Starlink. They had to switch to SpaceX for launch due to Russian shenanigans related to the Ukraine war. Russia said that they wouldn’t launch unless Britain pulled out of funding OneWeb, which they refused to do. Russia then basically held OneWeb’s satellites for ransom.
So OneWeb needed a new, reliable launch provider soon, and SpaceX was one of the beneficiaries.
Atlas SLV, Titan SLV, Saturn I and V (no failures), ‘Shuttle’ Space Transportation System, ‘Buran’ Soviet Shuttle, Proton (each variant), Titan I and II, MSLS, Minotaur (I/II and IV/V). Of course, you’ll note that all of these are either strategic weapon systems (or based on the propulsion elements of them) or were high value human rated systems, so they went through extensive development and qualification testing an have extensive independent mission assurance support compared to the “fuck around and find out” commercial approach. Each has their advantages; obviously no commercial company is going to build a couple dozen engines and test fire them to achieve qualification because of the enormous costs or run a qual campaign that tests the first ten articles to failure to understand capability, but then they’re going to experience a painful and often crippling learning curve. I’ve seen one commercial company after another claim that they’ve figured out the special sauce for instant high reliability and they won’t have any teething issues, and they all do regardless of how smart a crew they have designing and building them because launch vehicles are enormously complex systems with so many unforeseeable (or at least overlooked) failure modes.
As it happens, I’m currently trying to explain why trying to impose a requirement for “95% reliability at an 80% confidence level” for a vehicle with no successful flight history is a nonsensical requirement that could never be verified a priori even if some vehicle ended up demonstrating that with eventual flight history. A first order Bayes estimator, while underestimating failures of a commercial vehicle on first flight, is actually pretty good predictor of the number of failures of the first ten flights, and in general converges upon the realized reliability for a mature vehicle (although that is not really unexpected). On the other hand, trying to make predictions based upon Bernoulli distributions or from theoretical rollup analysis is so fraught with unvalidated assumptions that even when it produces something that incidentally correlates with realized reliability is clearly just coincidence because failures aren’t random or follow Gaussian-type distributions.
I haven’t gone through the full list, but most of those appear to be excluded by my conditions. The Atlas SLV wasn’t Convair/General Dynamic’s first rocket; that was (I think?) the SM-65 Atlas, which definitely failed its first flight. The Saturn V doesn’t count given the Saturn I, and the Saturn I seems to have a complicated history but it started with ARPA, which (I think?) had prior experience with other systems.
Since my question was probably confusing, let me rephrase a bit. I’m focused on the team here more than the rocket. Has a team of people with no prior orbital rocket experience ever succeeded in launching to orbit on the first try? Obviously this is a slightly fuzzy question since individuals can move around, and may themselves have varied background, but I want to exclude cases where some large group moved wholesale to some other organization. I’m looking for a “startup” that succeeded on the first try, even if that “startup” is embedded in some larger company or government group.
So, although the Atlas SLV is essentially the same body structural and propulsion elements as the Atlas ICBM, the avionics, guidance and control system, and forward structure/interface were all unique to the crewed launch vehicle. Similar for MSLS and Minotaur I/II to Minuteman, Minotaur IV/V to Peacekeeper, et cetera. Even though they may look externally similar there are enough differences in the systems that tend to fail that they can be considered essentially new vehicles (albeit with subsystems that are extensively qualified and demonstrated by flight experience). Saturn I and V are such different vehicles (and even the I and IB) that even though they share some common hardware they can be considered completely different systems. Even with a ‘common’ architecture, the addition of new flight critical systems (designated as “first flight items”) bring considerable attention and potential risks to an existing vehicle design.
That is a salient point because most of the failures that occur with a vehicle are not ‘random’ failures of a component that aren’t screenable but are failures of process and knowledge. Even mature launch systems like Proton and Soyuz can and have suffered failures because of the loss of critical knowledge and experience, resulting in a lack of understanding the implications of a design change or process ‘improvement’ with wide-ranging implications. To an extent, the experience and knowledge of the design and integration team are as important as the ostensible integrity of the design. That being said, every new design or major change is a challenge. Delta III was a massive fiasco even though it drew heritage and experience from the long-successful Delta II they made too many changes to basic architecture and tried to cut too many corners in pursuit of cost reduction. By comparison, the more capable Delta IV was highly successful, but also extremely expensive and not very competitive on the commercial market that Boeing had hoped to enter.
Regardless, space launch vehicles have a learning curve that can’t be short cut by doing some handwaving reliability analysis, and it can either be met by extensive qualification and system level testing along with extensive training and procedure development, or else accepting a high likelihood of failures in the first few flights until enough knowledge is gained to improve design and processes. I don’t think any group of space ingenues have ever successfully launched an orbital mission even though they come in believing they can figure out by just being smart. There is both a lot of entrenched knowledge and a lot of due diligence that go into making a launch successful, and it all ultimately comes from experience. And even that isn’t always enough to guarantee success.
I need to read up on Bayesian statistics. I’m a ‘payload’ engineer, and while my current work is based on statistical sampling methods, I never dig dig into the question of just how those numbers were generated. I’d like to better understand how newer development programs structure their development testing. Granted too, that one of my ‘sucessful’ payload system tests would be a major negative geopolitical event, we don’t need (or want) to go that route lest others wanna try to follow suit.
All that being said, I did get to see the SpaceX launch and recovery today, which was a bucket list item checked off. Admittedly, it was a little quieter than I expected . . . more of a low rumble than a loud roar I was expecting. The exhaust, even three miles away, was as bright as the sun and as long as the rocket itself. It looks like Relativity has been rescheduled for either the 11th, or next week. . .
Tripler
And true, @Stranger_On_A_Train , while it may be Florida, it’s better’n Afghanistan.
Agreed. I would add is “culture” as well, because it’s the culture that determines whether, for example, processes are taken seriously. And not just following processes (instead of ignoring them when they are inconvenient), but creating processes that can be followed, and also ensuring that the processes accomplish their basic intent. Since these processes often cross divisions, there needs to be a pervasive culture that treats them in a healthy manner. And that’s just one example of many that culture can affect.
It seems to me that you can’t just materialize a heathy organizational culture out of thin air, and that it must go through a “trial by fire” to some extent. And that so far, no fresh-out-of-the-box organization has successfully made orbit on their first flight. The successful ones looked at their failures and fixed not just the problems that arose, and not even the process gaps that led those those problems, but the cultural problems that led to the process gaps. Or, the cultural problems that led to the design problems, or the interface problems, or the information silo problems, etc.
And I did not mean to minimize the difficulties in bringing up even an iterative design; they surely are each their own challenge (especially in the presence of external factors, like politics). But there appears to be something extra challenging about that first one.
Excellent! Did you at least get the nice crackle of the exhaust sound? It’s something that is never quite captured properly by recordings.
And the exhaust as well–as you note, it is as bright as the sun. Maybe a good HDR display could capture some of that.
All of that is true; any team is composed of people, and regardless of assigned “roles and responsibilities” you really only understand their strengths and weaknesses over time. Systems engineering (which every organization large enough to develop and integrate a complex system like launch vehicle applies in some form regardless of what they call it) is supposed to help coordinate work and communicate issues but it is no substitute for real experience and connections that a team develops that makes a vehicle or mission successful or not.
The first vehicle is a complicated system for which there are just many conditions that cannot be tested in a fully integrated configuration on the ground, and can only be roughly approximated in analysis. It is also being developed at the time that a new organization is generally still formulating their processes and working through how to balance priorities. Once you have a certain level of knowledge—and you learn vastly more from your failures than you do successes—you know what to look for and how to avoid problems, both in the design of hardware and software, as well as organizational problems that prevent communication or effectively dealing with problems, but it is a steep curve to get there.
Interestingly, when you see previously reliable vehicles and mature organizations start having problems, it is not generally that they’ve just become complacent at the functional level; it is a combination of new people coming in who haven’t gone through those early failures, and are just doing exercises of analysis and testing by rote rather than being critical of potential issues, and upper management pressing to cut corners on tasks that they view as “low value” because they haven’t revealed problems in the past (or if they have, they are issues that are quickly resolved and presented). Eventually, you end up with a lot of cut & paste of testing procedures and cursory review of results, and a sort of cargo cult mentality toward dealing with indications of problems that just haven’t gone to failure yet. Those early, hungry days breed a diligent culture which atrophies with too much success and too many new people who haven’t been through it, and creates failures that would have been previously caught.
I was in Titusville, at a hotel catty-corner from the “American Police Hall of Fame”, and Ithink it was about three miles from the Kennedy Space Center’s visitor center–right across the bay from that. So, maybe ten miles, ish?
‘Taliban takes over Kabul.’ ‘DeSantis takes over Disney.’ Hrm, good point. . . If DeSantis, as POTUS, starts mandating the wear of burqas mullets, I’ll revolt.
Tripler
I told the SIGO we’ll be back at that hotel for vacation. I’ll bring my camera next time.
