Under the current 10 km flight program, all 3 Raptors are used at launch. During the ascent, they shut down one at a time.

After a flop and free fall of several minutes, they are reignited, with the best 2 used for landing.

Major issues with leaks and thrust control have emerged with tests SN 8 - 11, whereas the SN 5 and 6 constant light vertical flights went well.

Using a 10 km vertical attitude only (with parachute instead of flop) was proposed, as SN 5 and 6 did not have an engine out/relight in their flight sequence, as a "bridge step" to what they are trying with SN 15.

Could they try the engine out/relight (also vertically) program during the SN 15 static fire test?

It seems on the 2nd light, the fuel could be exhausted, giving the technicians a rocket with its critical components intact to look at after the test.

Could a more robust static fire program for SN 15 help it avoid its predecessors fate?


2 Answers 2


The static fire is just to make sure the engine will start, not to simulate the entire flight profile.

The problem isn't the engines themselves - they work just fine on ascent. The problem is that the rapid transition from horizontal to vertical creates all kinds of interesting stresses and flow dynamics in the tanks and the plumbing, leading to insufficient or inconsistent fuel pressure to the engines. This isn't something that can be easily sussed out on the ground, at least not at scale1.

They're going to blow up a few more prototypes before they figure this out. As long as the root cause isn't the same twice in a row (and they aren't raining flaming metal death on populated areas), then they're making progress.

  1. Yes, given time and money and space, they could build a test rig to suspend a full-scale Starship above the ground and rotate it from vertical to horizontal and study those flow dynamics with inert liquids. That would be an engineering challenge all on its own, it would take a non-trivial amount of time and money to build, and probably would not sufficiently capture all the corner cases that flight testing will.
  • 2
    $\begingroup$ Agree. As long as they keep hitting the pad and missing the city, they can go at it for as many tries as they want. $\endgroup$ Commented Apr 14, 2021 at 21:12
  • $\begingroup$ "Rapid transition from horizontal to vertical creates all kinds of interesting stresses and flow dynamics" Exactly! The external can handle it, the internal is having issues so: hang it from Tankzilla horizontally with a counter weight, then drop the weight and allow it to go vertical. "Study flow dynamics with inert liquids" That's step one cryro with nitrogen. $\endgroup$ Commented Apr 14, 2021 at 23:00
  • 2
    $\begingroup$ @Saiboogu: I think they lease the big cranes rather than own them outright. And the last thing you want to do with a crane is swing a load from it. That’s a good way to wreck a crane. $\endgroup$
    – John Bode
    Commented Apr 15, 2021 at 3:23
  • 1
    $\begingroup$ @AhmedTawfik I don't see it considering they don't actually want to reuse these things. They want to see how they perform under real conditions and then build another. What a lot of questions like this miss is that building a lot of starships is just as much of the point as anything else. They want to build 1000's of these. Every time they build one they are learning just as much about building (both the ship and the factory) as they are flying. $\endgroup$
    – eps
    Commented Apr 15, 2021 at 13:14
  • 2
    $\begingroup$ @RobertDiGiovanni SN15 is supposed to have a number of design improvements over SN8-11 (which is why they skipped 12, 13, and 14). And we have to remember they trashed a lot of F9 boosters when dialing in the landings for them. I fully expect them to succeed within the next couple of attempts. $\endgroup$
    – John Bode
    Commented Apr 15, 2021 at 13:50

SpaceX has both vertical and horizontal Raptor test stands, to test operation in both orientations.

The relight issues could come down to engine performance/quality -- It is a new engine entering mass production, so they have growing pains to work out. This is why they are A) building so many and B) testing them in flight.

The relight issues could also come down to the extreme motions of flight and the flip - this is difficult to simulate on the ground, and another reason to test in flight.

And the engine failures have eclipsed this portion of the test campaign in the public eye, but remember that the entire descent profile is new ground. They have a lot of new hardware in the flaps, and lots of interesting/tricky avionics work to fine tune controlling the rocket on the aerodynamic phase of ascent. This means they have other reason to fly, beyond simply nailing the landing. Each flight gathers more flight time for the avionics systems.

So, now that the engines work well enough on ground and they have a lot of flight testing to do, they are testing engines in that more challenging (and realistic) environment).

  • 6
    $\begingroup$ @RobertDiGiovanni Building many prototypes is not a downside. They intend to build and fly thousands of Starships and the operation in Boca Chica is every bit as much a prototype factory as it is a rocket prototyping facility. They do test the entire rocket vertically - they pressure test, cryo proof, and static fire them on the pad before flight. $\endgroup$
    – Saiboogu
    Commented Apr 14, 2021 at 17:18
  • 7
    $\begingroup$ @alephzero, 1960s-vintage engineering has certain advantages, such as rapid development. The F-111 (60s engineering) went from initial specification to series production in six years; the F-35 (21st century engineering) still hasn't entered series production after 26 years. I'm betting that 60s-style engineering is going to get Starship into orbit in less time and a lower budget than it's taking 21st-century engineering to launch SLS. $\endgroup$
    – Mark
    Commented Apr 15, 2021 at 2:26
  • 2
    $\begingroup$ @alephzero What throws away more money, bending metal and pushing it to destruction, tweaking to go further each time -- Or spending a decade or more modeling an ultraconservative design that likely won't ever fly due to the project lifetime exceeding the political cycle? I have my ideas, you may have yours .. But conveniently time will tell the difference soon enough. $\endgroup$
    – Saiboogu
    Commented Apr 15, 2021 at 3:12
  • 6
    $\begingroup$ They got to the moon in the 1960s. $\endgroup$
    – Robyn
    Commented Apr 15, 2021 at 3:26
  • 2
    $\begingroup$ @alephzero With 21-st century engineering, you mean not getting anything done at all, in decades? Because thats whats happening to most programs right now. I encourage you to read Wayne hales blog, he has some very interesting pieces about actually doing stuff instead of thinking about it, endlessly. SpaceX brings results. The 60s brought results. It might not be so bad to re-visit some older approaches. rapid prototyping is in fact a staple of modern software development. Yes, they throw away prototypes, but at the same time they progress, instead of just designing in theory. $\endgroup$
    – Polygnome
    Commented Apr 15, 2021 at 19:37

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.