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

[gongora]

Hopefully the failure mode for that engine was relatively simple otherwise there might be some effect to the DM2 schedule. I suppose NASA will be interested to know what went on as well.


Given how many of these have flown and how much firing time has been amassed this was probably due to either wear and tear or a manufacturing defect specific to that engine.

Considering that the Atlas engines that launch the Starliner end up in the Atlantic after one use, why would this delay Dragon 2 launch? This is the 5th launch of these engines and Dragon will not be on a 5th used booster. Talk about one rule for one...!

An engine failed on ascent.  That has to be looked at.

Yes it will be by SpaceX, NASA shouldn't be a part of the equation.

You can bet both NASA and DoD will be looking into it as they both have SpaceX launches in the next couple months.  An engine failure in flight draws attention, as it should, especially on a human rated vehicle that will be carrying humans for the first time soon.  That doesn't mean NASA or DoD should need to start their own investigations or create unnecessary delays but they would be negligent if they didn't at least take a serious look at it.

[abaddon]

So after Atlas's first flight NASA would have to go fishing. You are missing the point of my comment. It worked pefectly on its first flight and Dragon will be lauched on a new booster.

It failed when it wasn't supposed to.  Was it due to age?  refurb issue?  handling issue?  Something that could break the first time and SpaceX has just been lucky to date?  Who knows.  But to write:

NASA shouldn't be a part of the equation.

is just a big fat "not going to happen...NASA will be involved...full stop."

I have no idea if NASA will be involved or not, it is just my opinion that they shouldn't be. And I've given my reasons for that opinion.

Your opinion is objectively bad.  NASA has way too many eggs in the SpaceX basket not to expect and deserve to be treated as a very close partner.  And SpaceX, by all accounts, embraces that.  Why is it bad for NASA to be involved and have insight?

While we're at it, you can bet the USAF [EDIT: or I guess USSF) will have insight into the investigation as well.

[marsbase]

Of course government partners on future missions will be involved and consulted.  If this turns out to be the same engine that resulted in the abort on the previous attempt,  it would make me think that SpaceX took a calculated risk, knowing there might be a problem.  They might well have already decided that 5 times was enough for that booster and that they could afford to risk their own  payload in an attempt to maintain the launch cadence.   If all this is true, I'm sure NASA already knew about this risk assessment before today's launch.  SpaceX gained valuable data on the behavior of Falcon 9 in the event of an engine loss scenario.  And as Elon already said, this shows the value of having redundancy in engines.  I'm sure NASA and the USAF are following with great interest.

[thirtyone]

I agree that any sane customer (like NASA or DoD) will want to see the specifics of the engine failure. It's on a mission critical path.

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.

To be clear, I'm not saying anything is necessarily better than the other, just that this as an incredible wealth of data. This of course does not take into account whole vehicle reliability. If you ever really want to routinely fly humans, the best way to guarantee your safety modeling is correct is volume, so it's just kind of awesome that they have this much real world data.

[Rocket Science]

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.

Yes the braking/entry burn startup looked different. It is possible that the engine that failed was one of the three engines needed for this burn, so we only got a 2-engine braking burn, which could explain the apparent yaw and off-axis thrust before stabilizing, and the movement after cutoff. And that the speed was not reduced sufficiently, leading to loss of stage.

But I'm just speculating. The primary missing succeeded, yay for engine out capability.  (first time this has been demonstrated post F9v1.0)

Lars riding on the speculation train...

[CJ]

I think it's great that they lost an engine, yet still completed the actual mission. I also think it's a terrible loss that the recovery failed, because being able to examine that engine might have been very useful.

I do wonder if the engine with the problem was the same engine they had a sensor issue on.

 

[Lars-J]

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.

Yes the braking/entry burn startup looked different. It is possible that the engine that failed was one of the three engines needed for this burn, so we only got a 2-engine braking burn, which could explain the apparent yaw and off-axis thrust before stabilizing, and the movement after cutoff. And that the speed was not reduced sufficiently, leading to loss of stage.

But I'm just speculating. The primary missing succeeded, yay for engine out capability.  (first time this has been demonstrated post F9v1.0)

Lars riding on the speculation train...

Oh no, I guess pigs are flying... Or whatever the idiom is. 

[Rocket Science]

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.

Yes the braking/entry burn startup looked different. It is possible that the engine that failed was one of the three engines needed for this burn, so we only got a 2-engine braking burn, which could explain the apparent yaw and off-axis thrust before stabilizing, and the movement after cutoff. And that the speed was not reduced sufficiently, leading to loss of stage.

But I'm just speculating. The primary missing succeeded, yay for engine out capability.  (first time this has been demonstrated post F9v1.0)

Lars riding on the speculation train...

Oh no, I guess pigs are flying... Or whatever the idiom is. 

But the payload still made orbit with help of this "flying piggy"!

[CorvusCorax]

I agree that any sane customer (like NASA or DoD) will want to see the specifics of the engine failure. It's on a mission critical path.

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.

To be clear, I'm not saying anything is necessarily better than the other, just that this as an incredible wealth of data. This of course does not take into account whole vehicle reliability. If you ever really want to routinely fly humans, the best way to guarantee your safety modeling is correct is volume, so it's just kind of awesome that they have this much real world data.

Exactly that. This engine out scenario tells customers a lot about Falcon 9's reliability. In a good way!

A perfect launch doesn't give you much data about those low likelihood what-if scenarios that happen on a bad day. But some customers are obviously interested in these what-if scenarios - especially for high profile missions.

And if these what-if scenarios are proven to still result in a successful launch, even if booster engines start disintegrating in midair than this is a very good reason to fly on that booster in the future.
(Hypothetical scenario of course - we don't know the exact nature of the "shutdown" yet or how violent it was. There was a fireball in the plume visible from ground, but that could be nothing more than a bit of vented unburnt propellant flaring up in the exhaust)

As such, if I were NASA, DoD, or another high profile SpaceX customer, I'd be thirsting for the data and investigation results from this launch. And SpaceX can proudly share it (possibly with a non disclosure agreement on some details) because it comes with 1st order bragging rights!

Granted, "We lost an engine during ascent" doesn't sound good at first glance. But at second glance "We lost an engine during ascent on an extremely worn booster that was tortured beyond measure - and it still made a perfect orbit." Is a quality that customers like NASA or the military are very very interested in.

[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?

[rockets4life97]

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?

Yes, the Falcon 9 is nearly as impressive as the Space Shuttle! Thanks for reminding everyone.

[punder]

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?

Now compare their development costs and maintenance costs. And tell us what NASA plans to do with those magnificent jewels of technology in the next few years. Don't mean to be snarky, you're one of the most knowledgeable posters here, but...

[Lee Jay]

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?

Now compare their development costs and maintenance costs. And tell us what NASA plans to do with those magnificent jewels of technology in the next few years. Don't mean to be snarky, you're one of the most knowledgeable posters here, but...

But what?  The claim I answered was, "...the Merlin 1D engine has better statistical limits on its reliability than any other engine ever...".  That statement is false.

[marsbase]

But what?  The claim I answered was, "...the Merlin 1D engine has better statistical limits on its reliability than any other engine ever...".  That statement is false.

You are saying there were 405 shuttle engine-flights.  Not 405 engines built, if I understand you.  And thirtyone is saying there were 369 actual M1Ds built and flown at least once, not counting MVac engines, which are essentially the same thing.  He is not counting reuse.  If he counted reuse, the stats on Merlin would be much better.   

[su27k]

But what?  The claim I answered was, "...the Merlin 1D engine has better statistical limits on its reliability than any other engine ever...".  That statement is false.

You are saying there were 405 shuttle engine-flights.  Not 405 engines built, if I understand you.  And thirtyone is saying there were 369 actual M1Ds built and flown at least once, not counting MVac engines, which are essentially the same thing.  He is not counting reuse.  If he counted reuse, the stats on Merlin would be much better.

Yeah, there're probably around 85 core flights with M1D, so that's about 765 engines flown, some reused.

[punder]

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?

Now compare their development costs and maintenance costs. And tell us what NASA plans to do with those magnificent jewels of technology in the next few years. Don't mean to be snarky, you're one of the most knowledgeable posters here, but...

But what?  The claim I answered was, "...the Merlin 1D engine has better statistical limits on its reliability than any other engine ever...".  That statement is false.

To be similarly pedantic, you didn't "answer" anything. You asked pointedly rhetorical questions.

Big Picture is important. Merlin is cheap, robust, and efficient. SSME was horribly expensive and operated on the hairy edge every flight.

[Comga]

But what?  The claim I answered was, "...the Merlin 1D engine has better statistical limits on its reliability than any other engine ever...".  That statement is false.

You are saying there were 405 shuttle engine-flights.  Not 405 engines built, if I understand you.  And thirtyone is saying there were 369 actual M1Ds built and flown at least once, not counting MVac engines, which are essentially the same thing.  He is not counting reuse.  If he counted reuse, the stats on Merlin would be much better.   

My tally is that there have been 840 "flight engine cycles" since CRS-1 on October 7, 2012 excluding today.
Some of these "flight engine cycles" include static fire, first stage flight, boostback burn, entry burn, and landing burn, although some are just single firings of the second stage.
That's pretty good statistics.

(And contrary to statements made by others, the CRS-1 engine failure did not strand the experimental satellites in a bad orbit. 
NASA's rules required higher certainty than SpaceX had that boosting them from the Dragon deploy orbit would take them safely above the ISS.
The rocket was still capable of it.)


And most of us have great affection for the SSME's and admiration for their engineering, manufacturing, and maintenance.  Truly marvelous machines, but what they were not was part of a robust, economical system. 

[su27k]

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?

BTW, STS-51-F is not the only time SSME has in-flight anomaly, STS-93's SSME certainly didn't work correctly either and it also resulted in premature shutdown.

[Nomadd]

But what?  The claim I answered was, "...the Merlin 1D engine has better statistical limits on its reliability than any other engine ever...".  That statement is false.

You are saying there were 405 shuttle engine-flights.  Not 405 engines built, if I understand you.  And thirtyone is saying there were 369 actual M1Ds built and flown at least once, not counting MVac engines, which are essentially the same thing.  He is not counting reuse.  If he counted reuse, the stats on Merlin would be much better.   

My tally is that there have been 840 "flight engine cycles" since CRS-1 on October 7, 2012 excluding today.
Some of these "flight engine cycles" include static fire, first stage flight, boostback burn, entry burn, and landing burn, although some are just single firings of the second stage.
That's pretty good statistics.

(And contrary to statements made by others, the CRS-1 engine failure did not strand the experimental satellites in a bad orbit. 
NASA's rules required higher certainty than SpaceX had that boosting them from the Dragon deploy orbit would take them safely above the ISS.
The rocket was still capable of it.)


And most of us have great affection for the SSME's and admiration for their engineering, manufacturing, and maintenance.  Truly marvelous machines, but what they were not was part of a robust, economical system. 

Your tally might be a little high if you're talking about M1D reliability. The mission after CRS-1 was CRS-2, which still used Merlin 1C engines. Or did you exclude that one?
 My count now would be 78 M1D missions, including 3 Heavies, minus 1 US engine from CRS-7 which never got to fire, and the IFA US engine which didn't exist. I come up with 832 M1D uses.   Or 84 boosters plus 76 Vacs.

[penguin44]

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