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

[AndrewRG10]

Had the first stage safed the afts? If it hadn't reached that point perhaps the booster was terminated by the flight computer.

Not sure when it happens on downrange missions but RTLS flights save the AFTS right before entry burn. I'd guess same time or more likely even before. Probably didn't blow up due to the bombs onboard, I reckon it broke up on re-entry due to lack of transonic callout

[octavo]

I'm surprised that no one has yet mentioned what appeared to be a leak on S2?

I've never seen this before on a mission, and honestly it worried me far more than losing S1. It didn't seem to affect the final orbit, so obviously not too serious, and no one else has mentioned it, so I'm quite confused.

Does anyone know what this "leak" is? Multiple droplets form and roll down the S2 bell and pool together.

(See attached screenshot)

[EspenU]

I'm surprised that no one has yet mentioned what appeared to be a leak on S2?

I've never seen this before on a mission, and honestly it worried me far more than losing S1. It didn't seem to affect the final orbit, so obviously not too serious, and no one else has mentioned it, so I'm quite confused.

Does anyone know what this "leak" is? Multiple droplets form and roll down the S2 bell and pool together.

(See attached screenshot)

That looks like frozen oxygen from the vent lines. Happens on every launch and is completely normal.

[Vettedrmr]

From the update thread:

Here's a comparison between the telemetry for L4 and L5. You can see the premature M1D shutdown in the ochre L5 acceleration plot. By burning about a second longer, and not throttling down before MECO, the L5 booster makes up for the reduced thrust, and even separates about 16m/s faster than L4.

So, if I'm understanding this chart correctly, 1st stage burn didn't last longer, so the 1st stage was in a position to attempt landing.  Landings have to have everything go right to be successful, and based on the wonkiness of the re-entry burn and the lack of call outs afterwards, something went pretty wrong after TM was lost.

Or am I missing something?

Have a good one,
Mike

[AndrewRG10]

From the update thread:

Here's a comparison between the telemetry for L4 and L5. You can see the premature M1D shutdown in the ochre L5 acceleration plot. By burning about a second longer, and not throttling down before MECO, the L5 booster makes up for the reduced thrust, and even separates about 16m/s faster than L4.

So, if I'm understanding this chart correctly, 1st stage burn didn't last longer, so the 1st stage was in a position to attempt landing.  Landings have to have everything go right to be successful, and based on the wonkiness of the re-entry burn and the lack of call outs afterwards, something went pretty wrong after TM was lost.

Or am I missing something?

Have a good one,
Mike

Ascension burn and re-entry burn were same length, correct. However it didn't throttle down (or as much) leading to MECO. Could've used more propellent than planned, I don't know. It's just entirely likely the engine that failed was a landing critical engine. Caused bad entry burn and breakup.

[HVM]

Or like many have pointed out it's also possible that "shutdown" damaged octaweb's heat shield/airframe so that 1S was no longer stable enough for grid fins to control it, or heat shield integrity was lost.

[kdhilliard]

Had the first stage safed the afts? If it hadn't reached that point perhaps the booster was terminated by the flight computer.

You can hear the "Stage one FTS is safed" callout at T+6:52, but it is hard to make out as "Stage one entry burn start up" is called out at the same time.

https://youtube.com/watch?v=I4sMhHbHYXM&t=1300s

[Herb Schaltegger]

Probably didn't blow up due to the bombs onboard ...

Rocket flight termination systems are not “bombs.” There are linear shaped charges that run through the raceway that unzip the propellant tanks to allow prop mixing and combustion as aero forces break the vehicle apart into small pieces to minimize risk downrange from debris.

[ZChris13]

I'm surprised that no one has yet mentioned what appeared to be a leak on S2?

I've never seen this before on a mission, and honestly it worried me far more than losing S1. It didn't seem to affect the final orbit, so obviously not too serious, and no one else has mentioned it, so I'm quite confused.

Does anyone know what this "leak" is? Multiple droplets form and roll down the S2 bell and pool together.

(See attached screenshot)

I noticed this as well, although my first response was "hey cool, ice pooling on the gas generator exhaust duct, look at it bounce around"
Is that a common thing? I don't remember seeing that before.

[Eerie]

I'm surprised that no one has yet mentioned what appeared to be a leak on S2?

I've never seen this before on a mission, and honestly it worried me far more than losing S1. It didn't seem to affect the final orbit, so obviously not too serious, and no one else has mentioned it, so I'm quite confused.

Does anyone know what this "leak" is? Multiple droplets form and roll down the S2 bell and pool together.

(See attached screenshot)

I noticed this as well, although my first response was "hey cool, ice pooling on the gas generator exhaust duct, look at it bounce around"
Is that a common thing? I don't remember seeing that before.

I remember seeing it on every flight. Just check the videos again.

[baldusi]

The crazy thing to consider is that on a per-engine basis, the Merlin 1D engine has better statistical limits on its reliability than any other engine ever, simply because they launch so many per booster. Ignoring MVac, previous engine revisions, and reflights, there have been 369 produced and flown. The number of ignitions tested on an actual LV is ridiculous. Is this the first failure on a Merlin 1D?

For comparison, 116 RD-180s have been produced for Atlas III and Atlas V.

Weren't there 135 Shuttle flights, 3 engines each (many reused), with a single in-flight failure caused by two failed sensors?  And aren't those flights like 3 times longer than F9 booster flights?

51-L was an “engine” failure, too. And they had some very close near misses. And STS, a strict LV was equatorial LEO only. You should add IUS and Star48 reliability to the stack to do similar mission profiles.

[Tulse]

I haven't seen any talk about how the engine anomaly during the last Starlink launch might affect the scheduling of DM2. The announcement by NASA of media accreditation for DM2 seems to suggest there won't be any holdup, which I find a bit surprising.

[whitelancer64]

I haven't seen any talk about how the engine anomaly during the last Starlink launch might affect the scheduling of DM2. The announcement by NASA of media accreditation for DM2 seems to suggest there won't be any holdup, which I find a bit surprising.

I would guess that the approval for the press accreditation announcement for DM-2 predates the SpaceX Starlink launch.

[Tulse]

I would guess that the approval for the press accreditation announcement for DM-2 predates the SpaceX Starlink launch.

Probably, but given that it was actually publicly announced after the launch, I'm surprised they didn't hold it until they were sure there would be no impact.

[Lee Jay]

The crazy thing to consider is that on a per-engine basis, the Merlin 1D engine has better statistical limits on its reliability than any other engine ever, simply because they launch so many per booster. Ignoring MVac, previous engine revisions, and reflights, there have been 369 produced and flown. The number of ignitions tested on an actual LV is ridiculous. Is this the first failure on a Merlin 1D?

For comparison, 116 RD-180s have been produced for Atlas III and Atlas V.

Weren't there 135 Shuttle flights, 3 engines each (many reused), with a single in-flight failure caused by two failed sensors?  And aren't those flights like 3 times longer than F9 booster flights?

51-L was an “engine” failure, too. And they had some very close near misses. And STS, a strict LV was equatorial LEO only. You should add IUS and Star48 reliability to the stack to do similar mission profiles.

Obviously, I was talking about the SSMEs.

[TrevorMonty]

I would guess that the approval for the press accreditation announcement for DM-2 predates the SpaceX Starlink launch.

Probably, but given that it was actually publicly announced after the launch, I'm surprised they didn't hold it until they were sure there would be no impact.

Don't think engine failure will be issue given it has quite a few reflights before failing. Dragon 2 will only use new boosters.
The failure is not a bad thing for SpaceX especially if it is due to far wear and tear, will help them build more reliable engine.

[TheRadicalModerate]

I'm not confident enough to post this on twitter yet, but I think there might have been a second engine failure during the entry burn, and that killed the chance of landing.
 
At entry burn there's a change in the exhaust pattern accompanied by what looks like a bump, then the  camera got fogged up with visible drops of liquid, which I think is unburned fuel expelled during an engine failure.

F9 uses 3 engines for entry burn, if one of the engines that shut down on ascent was one of these 3 I don't think we'd see an entry burn attempt, and then if one of the engines used for landing failed during the entry burn then it might not even make it to the surface.

I saw the following weird stuff on the entry burn:

1) There's the piece of debris that comes off as they're adjusting the attitude before the burn.  Yeah, it could be ice, but then it's ice with square edges (see last 2 frames before it moves out of field).

2) Significant attitude deviation during the burn itself, with much higher than usual AoA.

3) Post-shutdown, the grid fins have a huge roll instability to damp out.

When you couple that with the hunk of debris visible during the pre-MECO failure, my working hypothesis is that there was a not-very-well-contained failure that damaged the skirt enough to cause aerodynamic problems once the core came down out of vacuum.  That's something that NASA is going to want to know about in some detail prior to DM-2.

I notice that we never got a call-out for landing burn start.  I'm guessing that it tumbled sometime after the reentry burn.

[Lars-J]

This amusing reddit posts is relevant: https://www.reddit.com/r/SpaceXLounge/comments/fkp8dp/in_todays_launch_of_spacexs_ice_there_is_a/

In today’s launch of SpaceX’s ICE there is a massive metal object. Can anyone tell what it is ?

 

[mulp]

From the update thread:

Here's a comparison between the telemetry for L4 and L5. You can see the premature M1D shutdown in the ochre L5 acceleration plot. By burning about a second longer, and not throttling down before MECO, the L5 booster makes up for the reduced thrust, and even separates about 16m/s faster than L4.

So, if I'm understanding this chart correctly, 1st stage burn didn't last longer, so the 1st stage was in a position to attempt landing.  Landings have to have everything go right to be successful, and based on the wonkiness of the re-entry burn and the lack of call outs afterwards, something went pretty wrong after TM was lost.

Or am I missing something?

Have a good one,
Mike

That's like saying things have gone "pretty wrong" when one toilet out of three in your house no longer seals after a flush and you must shutoff the water line refilling the tank,  and stop using iit, and depend on the other two.

Or maybe, a guest messes up one toilet during a party and now guests must be diverted to the other two toilets.

If you don't anticipate guests messing up a toilet during a big party, you totally failed at party design, and the party will be a total fail.

[mandrewa]

I haven't seen any talk about how the engine anomaly during the last Starlink launch might affect the scheduling of DM2. The announcement by NASA of media accreditation for DM2 seems to suggest there won't be any holdup, which I find a bit surprising.

I'm not going to do the math, but I suspect the Merlin 1D engine is the most reliable rocket engine (from a data perspective) ever.  And that of course depends on how you count it, and there are multiple methods that could be chosen to measure it, but I would try to make up a formula that counts both engine starts and time running, but with more emphasis on the number of engine starts than the seconds run.  I would also approach this with the assumption that every engine has an underlying likelihood of failure and that what we are trying to do is put bounds on what that underlying failure rate is.  Or in other words we should not assume perfection.  And instead that it is only through experience running an engine that we gradually raise our lower boundary on our estimate of the likelihood of failure.

The alternative is to assume that an engine that has never failed is perfect.  Not that there are many examples of that.

It seems a mistake to me to punish SpaceX for doing more testing and pushing their engines harder, which was in part to find problems, than anyone else has ever done.

They have already demonstrated a reliability that exceeds or is the same as the best of the alternatives.

It's always possible to do better, and rocket engines are a long way from beginning as reliable as other machines that we rely upon.  But I think that being the best in this context should be good enough.  And unless and until they actually identify a problem with the design that they can practically do something about, I don't think there should be a delay.