In SpaceX'S video, there is a conflagration inside the engine enclosure which seems irregular at around 1:49:49. The engine far left of the engine camera stops working and a few seconds later some plastic foil catches fire. Edit: As Damien correctly observed, that engine is restarted for the landing burn, so it is most likely not a failure, and certainly not a permanent one. Is this "conflagration" a side effect of a regular switching-off of the engine? Why would stuff burn inside the hull then?

From then on we also see what appears to be a leak on the bottom of the ship. White clouds emerge on the right side in the total view resp. bottom left on the engine camera.

Edit: The answers indicate that the engines were switched off intentionally. Is the "leak" some venting which becomes necessary for pressure relief after the fuel and oxygen flow from the tanks decreased?

If not, a leak could be the root cause for the "low tank pressure" leading to the eventual crash landing.

Was the flare and engine out intentional? Was the Starship supposed to fly (and land!) on two engines only after this point? Is the cloud normal?

  • 4
    $\begingroup$ "That engine never comes back on" - unless they for some reason choose to use a different camera inside the engine area during the landing sequence, it's one of the two engines used during the landing burn. $\endgroup$ Dec 10, 2020 at 8:08
  • $\begingroup$ @Damien_The_Unbeliever I would be a terrible witness -- you are right. $\endgroup$ Dec 10, 2020 at 9:23
  • 1
    $\begingroup$ Some stills: i.imgur.com/KER9nqc.png $\endgroup$
    – rkagerer
    Dec 10, 2020 at 12:02

2 Answers 2


I had the same question.

Elon Musk confirmed the "engines did great!"

There's also some discussion over here where the prevailing opinion is that the shutdowns were intentional.

Just before the engine on the left of the screen goes out, you see all three of them gimbal. The one being shut down moved out of the way, and the others took up a position allowing them to maintain similar overall direction of thrust. It happens again around T+3:12 when the second engine shuts down, but the effect is more obvious as the final engine takes up a central position.

In both cases, I think the gimbal activity included a short "kick" to impart a bit of rotational momentum away from the side of the engine being shut down, in anticipation of the brief "torque" you'll get in the opposite direction as the thrust transitions to the remaining ones (basically align the COM over the vector of the remaining engines).

In any event, the fact these steps took place before the flameout (if indeed it was a flameout) hints it was planned.

Scott Manley did a great commentary over the video and mentioned the event you described. He explains the fire was caused by gasses expelled from the shut down engine which got trapped in the enclosure then ignited.

  • 9
    $\begingroup$ Hopefully I didn't butcher the terminology too badly - I'm not a rocket engineer! $\endgroup$
    – rkagerer
    Dec 10, 2020 at 12:49
  • $\begingroup$ merely a scientist then? $\endgroup$ Dec 11, 2020 at 18:46

SE won't let me comment under or upvote rkagerer's comment. Yes engine shutdowns intentional due to the flight profile. Key points IMO:

  • given the minimum raptor thrust is around 40% or ~90 tonnes (https://twitter.com/elonmusk/status/1295553672454311941?lang=en), even at minimum throttle SN8 will still be accelerating upwards
  • SN8 must stop upward acceleration before the "belly-flop". The transition from vertical to horizontal must be controlled. If SN8 was to suddenly cut engines while still travelling upwards, it would continue on a "ballistic" trajectory (ie. still have upwards momentum) and lose aerodynamic control, so that when upwards velocity became zero and it was time to pitch over, it may have roll, yaw, be tumbling, etc.
  • it is clear from footage from various angles that the third and final engine actually kept SN8 in a hover and allowed it to move across land, or "translate" away from the pad. This was so after pitch-over, it could demonstrate that it could move across land as well as remain in a stable descent, to test out guidance to the landing pad (also allowed most of the descent to be over the ocean, where it was safer if engine re-light failed).
  • 1
    $\begingroup$ Thanks for the answer (no comment anyway). I watched Scott Manley's video mentioned in rkagerer's answer and wondered (again) about my bad observational skills. Small consolation is I was apparently not alone. The conflagration "inside the skirt", if that's the term, is normal (though burning off some foil is likely not). And yes, the transition from vectored thrust to an "aerodynamic fall" is delicate. One can see lots of cold gas attitude control puffs at the top during and immediately after shutting down the last engine. But after a few seconds it falls fast enough to make the fins work. $\endgroup$ Dec 11, 2020 at 6:22
  • $\begingroup$ And then what is the white smoke? $\endgroup$
    – Joe Jobs
    Dec 11, 2020 at 15:52
  • $\begingroup$ "SN8 must stop upward acceleration before the "belly-flop"." This should be obvious, no? Because otherwise it just keeps on accelerating upwards... $\endgroup$
    – RonJohn
    Dec 11, 2020 at 17:16
  • $\begingroup$ Aside - you now have enough reputation to do many basic things like comment etc stackoverflow.com/help/privileges $\endgroup$
    – Criggie
    Dec 17, 2020 at 19:42

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.