SpaceX Starship Orbital Flight Test

Seems like it is about as difficult to put together a focused concise video as it is to send a rocket into space.

The analysis (when we get there) seems to tie fairly well to Scott Manley’s observation that losing (say) 3 engines out of 33 has more impact than losing the 10% of engines might suggest. Presumably this is because much of the thrust from the remaining (say) 30 engines is in effect being used to offset the pull of gravity. The “net” thrust available for accelerating the vehicle thus suffers a larger degredation.

That brings us back to the speculation about why some of the engines were out. Did they fail to ignite? Or were they damaged by the Stage 0 flying debris during startup?


There hasn’t been any statement by SpaceX as to when the first three engines failed, and no video that shows the engines from below before the point at which SpaceX put up the status panel that showed them already out. Among those speculating, including me, the general consensus is that SpaceX would not have committed the ship to flight with any engines out, so it is more probable, especially how slowly the booster climbed from the pad and the fact that large chunks of debris were visible flying into the air around half the height of the entire stack, that debris struck the engines after launch commit and before the booster rose out of the zone of flying debris.

Another question is whether the engines that failed subsequently did so for other reasons or were also damaged by debris and simply took longer to fail than the first three. For example, a smaller piece of debris that punctured the regenerative cooling channels in the exhaust nozzle of an engine might cause an overheating problem which would take some time before the heat damage progressed to trigger an engine shutdown.


Given the repeated ‘gee whiz’ nature of Space X’s rocketry, what we already know of this failure of the relatively mundane launch platform and environs - the down-to-earth components of rocket science stuff - is quite surprising. If this turns out to be the root cause of the failure, I hope to hear a candid Musk explanation. I can’t imagine how this wasn’t foreseen.


It has been a worry and under discussion for some time. Here is Elon Musk in October, 2020.


This is his reaction after the launch.


The static test was conducted at 50% of full thrust and the damage was considered to be erosion and acceptable. This launch used (based on numerous reports, but nothing I’ve seen directly from SpaceX) 70% thrust, and the assumption was that damage would be acceptable. They were wrong.

Today’s (2023-04-27) post, “NASA’s Crawler-Transporters”, discusses the massive civil engineering project NASA had to undertake to build launch pads with flame diverters and trenches at a location in Florida where the water table was right beneath the surface. The situation at Boca Chica is almost certainly about the same, and SpaceX gambled they could avoid the massive pyramid-building-scale project with “FONDAG” (calcium aluminate refractory cement) and, later, a water-cooled steel plate.

This may have been an expensive misjudgment, and may cost much more time than iterating changes in the rocket. I’m sure they’re looking at lots of alternatives, especially since they now have large sunk costs in a launch facility in Florida which also assumes no need for a flame trench.


Are the hold downs locked until the thrust reaches 70%?


To my knowledge, nobody knows. There was one description of the launch sequence that implied that the hold-down clamps were released at engine ignition and the ship took off as soon as thrust was greater than its weight (which, as I understand it, is how Soyuz works), but I have never heard confirmation of this from any other source. It seems an odd way of doing business, since you’d usually want to be able to confirm you have all engines at full thrust before committing to flight. Falcon 9, for example, has had several cutoffs after engine ignition when engine monitoring saw something it didn’t like.


If so, there was no abort possibility if, for example, they discovered that the pad experiences excessive failure at 62%.


The acoustic shockwaves were visible in the first of the slow motion replays.

It’s entirely possible that those shockwaves, alone, were enough to cause some engine failures. That means (according to the conventional wisdom about shock dissipation) a “water cooled steel plate” of the kind shown in this video had better be able to disperse water droplets far enough above the plate to have the dissipative effect.

PS: I found this rather paradoxical result after looking up the extreme acoustic environment of missile silo launch:


Mundane stuff – there is a fascinating section in Eric Berger’s book “Liftoff” about the early days of SpaceX. The first Falcon 1 rocket launch failed in 2006. The cause was quickly traced back to “… the failure of an aluminum B-nut on the fuel pump inlet pressure transducer due to inter-granular corrosion cracking”.

That nut cost about $5 – a kingdom lost for want of a horse-shoe nail. The rocket had been exposed to corrosive sea salt spray while being prepared for launch from Kwajalein Atoll.

They had debated whether to go with aluminum or stainless steel. … Ultimately, Musk greenlighted the use of aluminum B-nuts, because they weigh one-third as much as steel. With rockets, every ounce matters”.

Engineering is always a matter of trade-offs. And bitter experience has often been the main way by which industries have learned which trade-offs are acceptable.


I meant this only in a relative sense, compared to the internal workings of the ship, itself. In professional negligence law, some of the decisions which turn out to be wrong are called “errors in judgment”. They are not negligence, provided the possibilities which turn out to have come to fruition were weighed and decided to be reasonable risks. Negligence would mean not having considered them at all. I imagine that is similar in engineering, so this will likely turn out to be an error of judgment, like the aluminum nut.


I am not sure if “error” is the best word – although we all have to accept that the law uses language in ways that have little connection with reality. The “error of judgment” is only apparent after the item has failed. Prior to that failure, it was simply the best engineering compromise that could be made with the information then available.

The underlying issue is that the only way to test certain concepts is to build them. The Tay Rail Bridge was cutting-edge engineering when it was built … and then collapsed in an unusually powerful storm in 1879. All subsequent bridges were better because of that painful learning lesson. The history of rocket launches is the same – a lot of spectacular failures leading to today’s highly successful launch industry.

With better information, we can make better engineering decisions. That does not mean the earlier decisions were “errors”. It is unfortunate that, in the real world, the only way to get that better information when working at the limits of prior experience sometimes involves pushing the envelope to the point of failure.


Study it forever and you’ll still wonder. Fly it once and you’ll know.
                                                        — Henry Spencer


Numerical simulation of a shock wave impacting a droplet using the adaptive waveletcollocation method

1 Like

Marcus House and Scott Manley, together again for the first time, dissect the aftermath of the Starship launch with Fraser Cain.


Mr. Musk on several occasions has made the point about the necessity of eliminating elements of a design wherever possible. Just because some element was included in a prior design does not mean it is still required for the current design. Presumably, SpaceX had a discussion at one time about whether a flame diverter would still be required – and a very convincing civil engineer demonstrated by sophisticated extrapolation from limited data that the answer was – No. But extrapolation is always risky, and now SpaceX has demonstrated that – Yes, a flame diverter is still required.

One important question is whether a suitable flame diverter system can be grafted onto the existing Stage 0, versus whether a completely redesigned/rebuilt Stage 0 foundation is required.

Another question is whether all the problems with the Starship flight stemmed from flying debris, versus whether there were additional problems with the rocket itself. Days, nights, & weeks of data analysis ahead for the SpaceX crew!

The bleeding edge of technology is a tough environment. SpaceX deserves a big round of applause for getting this first rocket off the ground. Let’s wish them well for the future.


One of the motivations for dispensing with a flame diverter and trench was that no such facilities exist on destinations such as the Moon and Mars, from which they will eventually want to launch return missions. If they can demonstrate that Starship (without Super Heavy, which is only needed for launches from Earth) can operate from unimproved terrain, that means they don’t have to find a solution for that chicken and egg problem. We now know that “size matters”, and while Falcon 9 can land with one engine on a flat concrete pad with no problems, a 33 engine Super Heavy won’t work. We’ll still have to see how Starship fares on its own.

NASA, in conjunction with Masten Space Systems, was researching a concept called “in-Flight Alumina Spray Technique” (FAST), in which a lander hovers above its landing location while injecting aluminium pellets into the rocket exhaust, where they are liquefied and sprayed onto the surface, creating an instant landing pad where the lander sets down. Unfortunately, on 2022-07-21, Masten filed for bankruptcy after running out of money trying to fulfill a NASA Commercial Lunar Payload Services contract on which they admitted to “grossly underbidding”. I don’t know if anybody else is pursuing the FAST concept. Here is a video about the idea.

One of the first things that will have to be done in establishing a Moon or Mars base will be constructing a landing site for subsequent landings with a surface that doesn’t spray debris in all directions and a berm surrounding it to retain any regolith that does go flying. A lunar lander can eject dust at velocities sufficient to enter lunar orbit and possibly damage spacecraft in orbit, not just sandblast anything in the vicinity.


Apparent owner:


Speaking of missile silos I first ran across the term “stage zero” while writing an AIA article for E’Prime Aerospace on the commercialization of the MX missile. The steam canister spit the missile up into the air which had two benefits: avoiding damage to the ground equipment which could be rapidly reused if necessary and imparting substantial momentum before the first stage booster took over.

For something like the Starship it would have to be quite a steam canister of course.


Fascinating! That is a real issue which will have to be addressed. Long ago, the Apollo missions dealt with this by having a two-part lander and leaving the lower part behind on the Moon, used as a launch pad. But that will presumably not be practical as a long-term solution for the Moon or Mars.

Suggesting launching from Earth without a flame diverter as a precursor to future launchings from the Moon & Mars is definitely the kind of “out of the box” thinking that would be desirable in a brainstorming session. But on analysis it would be found to be a non-starter – soils on Moon & Mars are different from Earth, as is gravity and atmosphere; even if it is possible to launch from an unimproved site on Earth, that does not necessarily mean it would also be feasible in the very different conditions on the Moon & Mars. Besides, no-one is suggesting launching from a totally unimproved site on Earth, not even SpaceX. If that really was a factor in the decision to dispense with a flame diverter, then I would be tempted to agree with CW’s characterization of the decision as an “error in judgement”. SpaceX had enough information at the time to make a better decision.

Just speculation: perhaps the major driver was government regulation. SpaceX was already heavily invested in Boca Chica when they began the planning for Stage 0. They may have realized that building a proper flame diverter would have triggered even more EPA delays because of the shallow ground water, and therefore decided to save time by trying to do without one. ???


Building a Saturn V style flame diverter and trench would have posed the same problems NASA faced at Merritt Island in Florida, where the water table was less than one half metre below the land. NASA built what amounted to a truncated pyramid 16 metres high at each of the launch pads with a ramp on one side that the crawler could ascend to deliver the launch platform and mobile service structure. The flame trench was excavated from the compacted soil of the pyramid and the flame diverter placed below where the launch platform was placed. This did the job for the Saturn V and later the Shuttle and SLS. Super Heavy, however, has twice the thrust of the Saturn V, so everything may have needed to be scaled up.

This would have been a massive civil engineering job and would have taken a long time just to compact the fill material to build up the pyramid. Since Boca Chica Wildlife Refuge borders the site, heaven knows how many forms would have to be filled out and hoops jumped through to build something like that.