15
$\begingroup$

What could be the reasons behind only 31 engines of the planned 33 engines igniting during Starship's recent static fire, given that the spacecraft is designed to have 33 engines and how could this impact the overall design of the spacecraft?

$\endgroup$
14
  • 4
    $\begingroup$ Elon Musk also mentioned that if it were a live launch, a loss of 2 engines would not be critical. Redundancy test could also be important in order to see how the rest of the system behaves when one or two engines have failed. A complex plumbing and propellant delivery could be a reason. $\endgroup$ Feb 20, 2023 at 13:36
  • 2
    $\begingroup$ @Moo - I did not quite get you. Musk did have a problem when his engine exploded. But why is a deadweight engine critical? $\endgroup$ Feb 21, 2023 at 4:24
  • 11
    $\begingroup$ The two engines did not fail "while running for 6 seconds", they failed at startup. Igniting a rocket motor is a pretty wild process. A running rocket motor is just a well running high performance device. $\endgroup$ Feb 21, 2023 at 13:17
  • 12
    $\begingroup$ One engine aborted while starting up, one was locked out by the test team from even attempting ignition. 31 engines ignited properly and ran for the full six second test duration. They were not having "two engines fail every six seconds". $\endgroup$ Feb 21, 2023 at 15:47
  • 3
    $\begingroup$ @Moo The most critical 6 seconds, yes? The 6 seconds during which engines are most likely to fail? Most plane crashes happen on takeoff or landing; you can't say that if 1% of planes crash during the first 60 seconds of flight, that means 60% crash during the first hour - it's more like if 1% crash within the first 60 seconds, then another 1% crash during the next 1-12 hours and another 1% crash within the last 15 minutes. (in reality it's less than 1%) $\endgroup$
    – user253751
    Feb 21, 2023 at 17:21

4 Answers 4

22
$\begingroup$

One engine was shut off because it had a leaky fitting. The other shut itself down (auto aborted) when it was fired.

The static fire probably wasn't delayed because 31 was thought to be good enough and they are on a tight schedule. They are very keen to get this launch done and delay is not what they want (more specifically it's not what Elon Musk wants).

31 engines was probably enough to give them and the FAA sufficient confidence. No doubt there will be more tests to come anyway. Starship should be able to complete its mission with a few engines out, especially if it's not carrying much payload.

Aborting the whole test would have required all the propellants to be recycled (with losses so might have required top ups), investigations and remedial work of undefined duration to be carried out, installation of one or perhaps two new engines (do they have appropriate spares immediately to hand at Starbase?) new road closures to be arranged and more nitrogen purges (might need more LN). The actual list of knock on things that cause issues in the background is probably a lot longer than that. For example I bet there is a long long list of changes and improvements to the GSE that would potentially be impacted.

$\endgroup$
6
  • 3
    $\begingroup$ How do you know that "specifically" it's not what Musk wants? $\endgroup$ Feb 20, 2023 at 14:27
  • 4
    $\begingroup$ I don't know for absolute certain, but given what I have seen and read about Elon Musk and from comments from people who have worked with him, it seems highly likely. He's a bit crazed about such things (I recommend Elon Musk's biography by Ashley Vance) $\endgroup$
    – Slarty
    Feb 20, 2023 at 15:15
  • 11
    $\begingroup$ @Slarty - Unlike NASA, Musk seems happy (well, not happy, but at least sanguine) for these things to blow up, as long as they're part of a learning curve $\endgroup$
    – Richard
    Feb 20, 2023 at 21:09
  • 2
    $\begingroup$ @Richard yes true, if it doesn't break your not testing it hard enough. Musk has an unusual attitude to risk that was probably at least in part instilled by his grim childhood experiences in apartheid South Africa. $\endgroup$
    – Slarty
    Feb 20, 2023 at 22:44
  • 6
    $\begingroup$ @Richard: That seems more a matter of an industrial approach. SpaceX plans to build rockets in volume, to keep costs down. This is very much the opposite of SLS. It keeps marginal costs down; losing an early-model Starship isn't that bad. $\endgroup$
    – MSalters
    Feb 21, 2023 at 13:31
20
$\begingroup$

Note that this test gave them data on firing all engine position "types" (gimbaled center cluster and inner ring, fixed outer ring), the effects on the booster and ground equipment of attempting to start up 33 (later changed to 32) engines, and operating conditions with all but 2 engines operating, with 94% of the originally planned thrust and 32 (planned) engine ignitions. All they really missed out on was a test fire of the engine that shut down in startup and the one that they didn't fire, and the first of those would most likely have happened anyway even if they did abort the test to fix the other engine.

Proceeding with the test cost them only the propellant they already had loaded, and they got almost everything they possibly could get out of the test, so in cost/benefit terms the choice is clear.

$\endgroup$
7
  • $\begingroup$ Note that the planned thrust was "~50% throttle". Launch is going to be...loud... $\endgroup$ Feb 21, 2023 at 15:45
  • 3
    $\begingroup$ This answers the OP's second question, but it doesn't answer the first. Why were the two engines shut down? Suggesting that it was only a test of engine types would further suggest that the two engines were never meant to be fired, but if that's the case, why shut them down when it would make more sense to never plan to fire them in the first place? $\endgroup$
    – JBH
    Feb 21, 2023 at 16:03
  • 1
    $\begingroup$ @JBH One engine was manually disabled (probably some issue was detected) prior to testing, an other engine was automatically disabled by the computer on startup... Since SpaceX does not have time travel technology there is no way that they could have known of the automatic shutdown in advance and so they failed to not ignite it. $\endgroup$
    – GACy20
    Feb 22, 2023 at 8:21
  • 1
    $\begingroup$ @ChristopherJamesHuff From the answer, "this test gave them data on firing all engine position 'types'." I apologize for not being more clear. My point was that the way the answer reads, it's unecessary to know why the engines failed simply because enough data was obtained that it didn't matter - except that it matters to the OP, which is the point of the question. $\endgroup$
    – JBH
    Feb 24, 2023 at 2:44
  • 1
    $\begingroup$ That was one of the things the test got them. And no, I didn't answer why the engine (singular) that shut down did so, and I'm not going to...as far as I'm aware, that information simply isn't available outside of SpaceX, NASA, and possibly a few people at the FAA. If the OP has an issue with a partial answer due to that, they can speak up themselves. $\endgroup$ Feb 24, 2023 at 4:10
15
$\begingroup$

In general, the materials cost of getting a large system like that ready for test is much less than the people-and-infrastructure cost of setting up the test event.

At that point, they had already put in the people-and-infrastructure cost. The real question was likely not whether they could have the test that they wanted (because the answer was to that was no), but whether it was worth paying the relatively small marginal cost in order to get a large subset of the data they wanted back.

I don't know what SpaceX's specific costs look like here. But from working on other large-scale projects, late delays in test events mean you have already committed most of your cost for that event. Picking a path forward is a balancing act between the high costs of pausing or restarting the test and the value of the data that you might miss from things left out in last-minute changes.

$\endgroup$
2
  • 2
    $\begingroup$ @Starshipisgoforlaunch it's lovely that you've edited literally every answer to this question, some more than once, but "the cost of propellant" you've added is part of the "materials cost" fectin was comparing the "people-and-infrastructure" cost to. You've made the answer less correct. $\endgroup$
    – Erin Anne
    Mar 1, 2023 at 10:51
  • $\begingroup$ Okay. I agree @ErinAnne $\endgroup$ Mar 1, 2023 at 11:21
11
$\begingroup$

Seeing how a system responds with an error (1 engine not fired, 1 aborted) actually provides useful information as well. For example, it shows that the software and hardware in place to shut down an individual engine works properly - otherwise an engine problem could easily spread to other engines and possibly cause loss of the spacecraft.

In addition, working through engine failure is designed into these types of rockets. For example, at least one time Falcon 9 launched successfully despite failure of an engine. In that instance, the first stage was not successfully recovered, but the primary mission succeeded.

On the other hand, if you only have one engine then failure of that engine is failure of the mission.

$\endgroup$
4
  • 4
    $\begingroup$ At one time I worked on an autonomous vehicle (for a university competition, not Tesla). Some time during testing something caused the control loop to freeze for a second and the robot ran head-first at full speed into a literal brick wall. Of course, the most expensive component (LIDAR) was at the front. Never found out why that was, but after it happened again, we added a secondary control system that would stop the robot if it sensed it was closing in too fast on a wall. I'm not sure what this has to do with rockets, but failed tests can be informative. One of the competitors' caught fire. $\endgroup$
    – user253751
    Feb 21, 2023 at 22:26
  • 1
    $\begingroup$ @user253751 The old autonomous vehicle competitions were fascinating looks into why the problem was so difficult and how many approaches you can take to try to solve it $\endgroup$
    – Kevin
    Feb 22, 2023 at 23:42
  • $\begingroup$ @Kevin btw this was like a 40cm cube-ish robot designed to move 10cm foam cubes for a competition, nothing like a car, if that's what you were thinking. $\endgroup$
    – user253751
    Feb 22, 2023 at 23:54
  • $\begingroup$ @user253751 Makes sense, they also did full scale actual car races through mountain roads and stuff, but those were large scale productions, not really college projects $\endgroup$
    – Kevin
    Feb 23, 2023 at 16:56

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.