SpaceX F9 : Starlink 6 (v1.0 L5) : Mar. 18, 2020 - Discussion

[wrvn]

I hope it wasn't the same engine that was out of family during the abort.

Ohh damn I really hope its not engine failure with Crew Demo so close 

What are you trying to say? I can't read your comment well.

I'm trying to imply likely Crew Demo delay if this was an engine problem.

[toruonu]

I'm trying to imply likely Crew Demo delay if this was engine problem.

It's confirmed to have been an early engine shutdown on ascent.

[AndrewRG10]

The last Starlink flight had exact same profile so we can analyse them and see if something is different.

Max first stage speed,
Starlink L4: 7861km/h 68.8km
Starlink L5: 7920km/h 69.5km

Both re-entry burns were 21 seconds and were within 1 second of each other.

Does this mean something engine-wise failed? Could it be just differences in atmosphere causing a margin of error that's expected?

The first 3 with the same profile had cut off speeds of 8040km/h, 8067km/h, 8080km/h (not in order of flights) 40km/h range in speed. This one had 60km/h difference so it's entirely possible to be just atmospheric differences. But the first 3 had no more than couple hundred metres difference in height. Who knows right now?

[wrvn]

they said it was a 10sec reentry. My counting reached 17, not very precise but It would be interesting to go back and time it properly.

So it looks like F9 tried to compensate for engine loss by extending entry burn. Maybe indicating that it was not center engine that failed.

[haywoodfloyd]

    Yeah. There was also an early engine shutdown on ascent, but it didn’t affect orbit insertion. Shows value of having 9 engines! Thorough investigation needed before next mission.
    — Elon Musk (@elonmusk) March 18, 2020

[space_snap828]

If a F9 is going to fail, a good place for this to happen is in booster recovery. 

Yeah, but indications so far point to recovery failing due to a Significant Event just prior to MECO on ascent so...

For example if an engine blew, it would cut down thrust for a bit if/until other engines throttled up to maintain expected acceleration at that point in flight. But due to that delay, the booster will have traveled further downrange prior to MECO (as it probably targets a specific MECO velocity) and hence it could have overshot the ASDS sufficiently that even if the problematic engine was not one of the 3 required for reentry and landing, it would still miss the zone as it would have put it beyond the possible landing envelope.

I have wondered in the past what limits there are to recovery, for the reasons you mentioned. Particularly as there is value to recovering a rocket after a problem like engine shutdown/explosion so they can physically investigate the problem. Here, if they could have landed, they could have actually seen what caused the engine failure.

[ugordan]

I have wondered in the past what limits there are to recovery, for the reasons you mentioned.

Well, it's not exactly trivial to choreograph all these things, determine at what state vector (velocity, height) to cut engines on ascent even on a normal launch as you'll always get some dispersions due to wind, steering losses, etc. At the end of the day your primary mission takes precedence which is why I think they program MECO not for most favorable landing state vector, but on mission requirement, i.e. target MECO velocity and/or altitude to satisfy S2 performance margin.

[EspenU]

they said it was a 10sec reentry. My counting reached 17, not very precise but It would be interesting to go back and time it properly.

So it looks like F9 tried to compensate for engine loss by extending entry burn. Maybe indicating that it was not center engine that failed.

To me it looked like an outer engine that failed. Also, if it was one of the 2 outer ones that are used for entry (as it looks like due to the longer burn), that might also explain the roll oscillation at engine cutoff.

[1]

I'm actually somewhat (irrationally?) pleased to see this. F9 has always been advertised as having engine-out capability; and we just saw that today with a successful Starlink deployment. The booster was lost, but not before putting its fifth payload where it needed to go. Congrats to B1048 for scoring SpaceX's first "ace."

[AndrewRG10]

Fairing catches have very likely failed. Been too long with no word as the SpaceX twitter account doesn't like reporting failures Hopefully they didn't hit the deck like last time and splashed soft for reuse.

[lrk]

I'm actually somewhat (irrationally?) pleased to see this. F9 has always been advertised as having engine-out capability; and we just saw that today with a successful Starlink deployment. The booster was lost, but not before putting its fifth payload where it needed to go. Congrats to B1048 for scoring SpaceX's first "ace."

That capability has been demonstrated before, but not since CRS-1 in the F9v1.0 days.  Even then the early shutdown lead to a incorrect orbit for a secondary payload, since at the time F9 did not reserve propellant for landing. 

[soyuzu]

Happened on ascent not on entry.  Wonder if it's the same engine that gave them the abort. Hindsight says should have  changed it out if yes.

https://twitter.com/jeff_foust/status/1240263546732240897

It’s possible that the cost and schedule requirements leans toward risking one engine-out result in a successful mission and landing failure.

[niwax]

The last in-flight engine shutdown was during CRS 1. Since then, there have been 79 successful flights. Depending on the precise number of static fires and landing burns, this puts reliability since then at somewhere around 1 in 2500 ignitions and 1 in 40h.

This particular engine had 11 - 26 ignitions and 13 - 17 minutes of operation, depending on which engine it is.

[cscott]

I'm actually somewhat (irrationally?) pleased to see this. F9 has always been advertised as having engine-out capability; and we just saw that today with a successful Starlink deployment. The booster was lost, but not before putting its fifth payload where it needed to go. Congrats to B1048 for scoring SpaceX's first "ace."

And even though failure happened shortly before MECO it implies the engines were in a steady state after the failure, as they managed to restart and complete an entry burn.  So the blast containment shields appear to have done their job and there was no collateral damage to other engines.

It will be interesting to learn if the landing failure was due to the failed engine being one of "the three" or if it was just trajectory-related, due to the extended compensation burn.  If it's a trajectory problem, it's possible that software or other tweaks (reserving more fuel, stationing a third ASDS further down range) *could* save such a situation in the future... although the cure might cost too much to be worth it.  Better to crank up the engine reliability further instead.

[kessdawg]

I'm actually somewhat (irrationally?) pleased to see this. F9 has always been advertised as having engine-out capability; and we just saw that today with a successful Starlink deployment. The booster was lost, but not before putting its fifth payload where it needed to go. Congrats to B1048 for scoring SpaceX's first "ace."

And even though failure happened shortly before MECO it implies the engines were in a steady state after the failure, as they managed to restart and complete an entry burn.  So the blast containment shields appear to have done their job and there was no collateral damage to other engines.

It will be interesting to learn if the landing failure was due to the failed engine being one of "the three" or if it was just trajectory-related, due to the extended compensation burn.  If it's a trajectory problem, it's possible that software or other tweaks (reserving more fuel, stationing a third ASDS further down range) *could* save such a situation in the future... although the cure might cost too much to be worth it.  Better to crank up the engine reliability further instead.

You're assuming there was a blast to contain. The engine could have just reached some shutdown criteria like high/low temp or power and shutdown before there was further damage.

It even could have been another sensor issue.

[ames]

You're assuming there was a blast to contain. The engine could have just reached some shutdown criteria like high/low temp or power and shutdown before there was further damage.

It even could have been another sensor issue.

The debris observed before the re-entry burn may be significant.

[kessdawg]

You're assuming there was a blast to contain. The engine could have just reached some shutdown criteria like high/low temp or power and shutdown before there was further damage.

It even could have been another sensor issue.

The debris observed before the re-entry burn may be significant.

Right, forgot about that.

[XenIneX]

So, looking at previous Starlink profiles, the entry burn duration for Starlink-6 seems typical.  The video for Starlink-5 dropped out, but the callouts were ~20 seconds apart; similarly, despite their different profile, Starlink 2-4 also had ~20 second burns.  Since the duration didn't substantially change, I'm going to assume that either the problem engine wasn't one of the landing engines, or it operated nominally for the entry burn.

I did notice that the F9 did heel over substantially as soon as the engines stabilized after relight, in a way that previous Starlink missions didn't (t+6:59).  That indicates, to me, that it was scrubbing excess horizontal velocity.  Then, after entry burn shutdown, it heeled over the other way (oscillating a bit, possibly due to the gridfins having a bit too much roll control authority) to guide its new trajectory.


Speculative event timeline:
Engine failure at t+2:21.  Booster extends burn to compensate, completing mission, but lofting the stage past its target.  Entry burn successfully scrubs excess horizontal velocity.  Resulting entry trajectory is outside the flight envelope, or too spicy to stick the landing.

[woods170]

I'm actually somewhat (irrationally?) pleased to see this. F9 has always been advertised as having engine-out capability; and we just saw that today with a successful Starlink deployment. The booster was lost, but not before putting its fifth payload where it needed to go. Congrats to B1048 for scoring SpaceX's first "ace."

Emphasis mine.

Engine out capability of F9 was already demonstrated 7.5 years ago, on 7 October 2012, during the CRS-1 launch.
During that launch, one of the nine Merlin engines blew it's LOX dome and the engine was shutdown roughly 80 seconds into the launch. The primary payload - a cargo Dragon - was successfully delivered into orbit.

[Citabria]

The flash appeared left of center, and the 5 visible plumes looked normal afterward, so maybe it was the lower left engine (seen from the camera POV).