SpaceX misses landing on a drone ship, breaking the company's streak | Hacker News Reader

SpaceX misses landing on a drone ship, breaking the company's streak

mashable.com

121 pointsrezist80871 comments

Sounds about right, given what Elon was expecting back in January [1]

> My best guess for 2016: ~70% landing success rate (so still a few more RUDs to go), then hopefully improving to ~90% in 2017

[1] https://twitter.com/elonmusk/status/689299216607232000

As the presenters repeatedly said throughout the livestream, even when the landing fails it is a great source of data to integrate into future landing attempts.

So far it doesn't appear that the landings have failed for the same reason twice. Not learning from past mistakes is really the only failure.

>As the presenters repeatedly said throughout the livestream, even when the landing fails it is a great source of data to integrate into future landing attempts.

That's sort of half joke and half spin. It's hard to argue having a failure and learning you shouldn't have designed your rocket with a particular flaw is better than not designing your rocket with that flaw in the first place.

Oddly enough, sometimes letting things fail is the cheapest way to create a better design. Not necessarily in direct costs, but getting a great product out in 30 years is rarely an option.

Has SpaceX actually cleaned out and reused any of their recovered rockets yet? I'm guessing no, they're still analysing them, but it'll be interesting when they do. (Damned if I'd want my payload on the first recycled rocket...)

Elon says first reflight's targeted for Sept/Oct. https://twitter.com/elonmusk/status/740296489532948480

remember to always take Elon's time preditions with caution, he originally said CRS8's booster (the first landing on the ASDS) would be ready for flight in "june, maybe july"

I'm sure they're doing their best, but space takes time.

"space takes time" is pretty good. Understood they could miss the date, but that's what Elon said they're targeting. If the first re-use is October 2017, I'll still be impressed as I'll bet the timeframes compress over time.

A re-used rocket should be more reliable than the first go-around. Purely because if anything is likely to go wrong it will either happen on the first run or after N runs that is enough to wear a component out. N is likely much greater than 1.

That may be true in the long term, but that may not be true for the first few attempts until they understand the sort of things that are stressed/weakened/generally affected by the first launch.

That's the hope, but we don't have enough experience to know whether or not it's true. The analogy people like to use is airliners, but airliners don't get anything like the g-forces or heat load you see on a returning rocket stage.

unless some significant wearing happens during descent

Unknown unknowns? As much as you prepare there will always be things that slip between the cracks.

Of course it is hard to argue that, It is much better learning that you shouldn't have designed your rocket with a particular flaw than the alternative.

The alternative is not designing your rocket with the flaw in the first place, using only known and proven technologies, avoiding risks at all cost like gobertment(NASA, ESA) agencies do.

And this is very costly in the long term, because technology does not improve.

s/flaw/optimisation

Sometimes the only way to improve a design is by accepting the chance if a higher failure rate.

Local maxima!

The failure is also the failure

Stabilized version here: https://www.youtube.com/watch?v=p5_hvVbxAAo

Till the very end it seemed that is almost made it.

From the little I could find on it, it sounds like the rocket actually landed... and then sorta kept landing until RUD. Is that right? There was more slam than hover in this particular maneuver?

That's pretty much correct. One of the engines wasn't producing full thrust during the final landing burn. Since there's almost no fuel margin, the rocket ignites at the last possible moment with the goal of reaching zero speed right as it touches the ship. Reduced thrust means that it would touch the ship before it had fully braked.

The live video from the ship didn't work well (something about satellite dishes makes them work poorly when there's a massive rocket blasting them with fire, apparently) but from the few frames that made it through, the rocket clearly managed to come to rest upright on the ship. There was a great deal of residual fire, everything was clouded in smoke and compression artifacts, and the last frame was from only about four seconds after touchdown. It's hard to tell exactly what was going on, but it looks like it hit hard, broke the legs or the fuel tank, and after a few seconds it exploded or fell over or both.

I'm somehow reminded of the Monty Python line about building castles in a swamp:

> When I first came here, this was all swamp. Everyone said I was daft to build a castle on a swamp, but I built in all the same, just to show them. It sank into the swamp. So I built a second one. And that one sank into the swamp. So I built a third. That burned down, fell over, and then sank into the swamp. But the fourth one stayed up. And that’s what you’re going to get, son, the strongest castle in all of England.

SpaceX may be doing things in a slightly different order, but maybe eventually all of their castles, er, rockets are going to stay up on the landing pad. :-)

That's hilarious, and a great comparison too. They'll get there! Fortunately they've figured out how to get paid even if they fail, so they can take their time getting it to work.

There is a satellite photo from 40min after the landing [1] and it looks like its still standing... So it might just have crushed the engines.

[1] https://twitter.com/deimosimaging/status/743153542362439680

There's some question as to what that actually shows. Here's one theory that shows it fallen over: https://twitter.com/NASASpaceflight/status/74318210287573811...

Reference: https://twitter.com/elonmusk/status/743097337782763521

"Looks like thrust was low on 1 of 3 landing engines. High g landings v sensitive to all engines operating at max."

Another thing to keep in mind this is another one of the launches where the 1st stage was coming in a lot faster (geosync orbit launch).

Update: they're now saying that it ran out of propellant before touchdown. Same basic result, but different cause.

At least it's a policy that ensures only very well-performing vehicles are reused.

Once all the weak rockets crash, we can take the surviving rockets and cross-breed them to build the ideal rocket race.

When I heard they were using genetic algorithms to improve the rocket's design, that's not what I had expected.

Truly, we are living in 2 billion years ago.

I don't care if they're good, I want 'em lucky!

I read a book involving that once. People had been bred using a lottery system over generations, this made them incredibly lucky.

Ringworld? I don't quite remember.

Yes, Ringworld.

Video is up: https://twitter.com/elonmusk/status/743602894226653184

Accordng to Elon, the rocket experienced #RUD = Rapid Unscheduled Disassembly

The rapid is kinda redundant, because if it's unscheduled, it's going to be caused by a spontaneous release of energy stored onboard (as opposed to a scheduled, planned disassembly, where the energy is applied by external workers). Most spontaneous releases of enough energy to disassemble a launcher will be rapid.

Of course, RUD is already established in the space industry's jargon.

RUD = it turned into "rudere" [1]

[1] https://en.wiktionary.org/wiki/rudere#Italian

https://twitter.com/elonmusk/status/743097668725940225

While this is bad news, even if they only land a quarter or a half of the time, it must still save a huge amount on launch costs.

How so? They aren't reusing the rockets...we don't even know if they are in a reusable state.

[SpaceX's Musk: We'll reuse today's Falcon 9 rocket within 2 months • The Register](http://www.theregister.co.uk/2016/04/09/falcon_9_rocket_reus...)

It's been 2.5 months. Last week, it was pushed back to the fall: http://www.space.com/33145-spacex-to-refly-falcon-9-rocket-f...

It is interesting that it was 2 months.. then after landing a few, it's now 6 months. Makes you wonder how much damage the rocket is sustaining during launch/recovery.

If I had to guess, their first priority is launching new rockets that generate revenue, with paying customers' payloads on them (as all of the recent Falcon 9 launches have been), rather than a self funded R&D flight. I don't think any paying customer would put anything other than a boilerplate/test article satellite on the very first ever re-use of a first stage.

People would if the cost was sufficiently defrayed -- this is just a financial instrument problem.

That might make the first reuse flight unprofitable for SpaceX, but not as unprofitable as one with no payload.

One of the biggest risks to SpaceX's business is probably the perception that they cannot launch on time, that they have multi-year delays: they are not reliable in terms of launch date expectations. While SpaceX is getting better at this, they need to be able to show prospective customers that they will do everything possible to launch on time, or with the fewest delays. Otherwise, cheaper launches be damned, many customers might switch to a launch provider that will commit to a date.

See this here for a series of graphs of how big their launch backlog has grown over the years: http://www.americaspace.com/?p=93456

This means that launching customer payloads on fresh-built rockets is Priority #1, at least until confidence is fully restored, cadence is up, and backlog is down.

The insurance companies which insure commercial telecom satellite launches are EXTREMELY risk averse - I would be surprised if you see a commercial launch in any of the first eight re-uses of first stages. Government/science payload/R&D, yes. LEO or geostationary transfer orbit commercial, no.

edit: you might see a commercial flight with a never-flown-before prototype commercial satellite bus (example: like a geostationary bus the size of a boeing 702 with a novel never before tested type of ion propulsion stationkeeping system, or a new battery system, or something else weird).

As I said, a financial instrument problem. If nobody is willing to create the appropriate insurance because of structural inefficiencies in the insurance industry, well, that's reality. But there's no prima facia reason why it should be uninsurable.

For context, this launch cost around $60mil and carried $200mil of satellite (spacenews http://spacenews.com/spacex-successfully-launches-2nd-pair-o... ) Spacex are aspiring to a 30% discount https://spaceflightnow.com/2016/03/31/spacex-hopes-to-sell-u... Until the tech is well proven this would seem to only be a compelling offer for lower cost sats. Theres also the catch 22 that the first mission will be almost uninsurable. I reckon we'll see an LEO demonstration launch maybe with some cubesats n the first one.

From what I understand the first "used" mission will be insured.

Insurers seem to be playing a bit loose with SpaceX:

>This insurance official said that given the market’s current softness, underwriters have overlooked the fact that they are not always sure what modifications have been made to the Falcon 9 they are insuring.

http://spacenews.com/spacex-to-brief-underwriters-on-the-roa...

If, say, around 30% of rockets blow up, I don't see why there wouldn't be an insurance willing to write a policy for say 35%?

Not if the satellite is much more expensive than the rocket. Using the GP's figures of $200mm and $60mm, that would mean paying $70mm in insurance instead of just waiting for the next $60mm flight.

Yes, that's true. With an expensive enough satellite, even a free launch would not compensate for 30% failure.

Supposedly SES volunteered to put their next flight on a used stage. The sticking point was price - SpaceX wants to charge $40m and SES wanted to pay only $30m.

SpaceX claims several customers have volunteered to be first. I don't think we're going to see a test payload unless SpaceX is concerned there's such a high chance of failure the PR hit would be to big to risk it.

One apparent problem is also to get a rocket certified for flying a second time. No one did that so far so there are no procedures for it. Part of the delay is figuring that out.

They want to do their first relaunch with a paying customer -- meaning the customer's insurance needs to be willing to insure a relaunch. So, now, they're talking to insurance companies to see what will convince them that the risks of a relaunch are comparable to those of a new rocket launch. This may involve doing a teardown on one of the returned stages, just to check for damage after the first launch and landing, even though they do not expect to do anything like that level of work on stages that they actually relaunch.

An article that goes through all this in excruciating detail is here: http://spaceflightnow.com/2016/06/03/spacexs-latest-booster-...

For the record, it looks like Musk never said "we will do it in two months". He said SpaceX would aim for June and even mentioned likely delays in the same sentence. There's a huge difference, and the headline you linked to was written as clickbait.

More recently:

"Aiming for first reflight in Sept/Oct." as of June 7th.

https://twitter.com/elonmusk/status/740296489532948480

That's musk time though so you need to multiply by 2.

...more than 2 months ago...

They say that even the "max damage" stage could fly again:

https://twitter.com/elonmusk/status/734274360588926976

If this was successful the title would have read "Elon Musk breaks record for most landings on drone ships."

I figure every success is icing on the cake until they drop prices on the assumption they will always recover the rocket

RUD, Well, a wave hit it.

Losing 1 in 4 is a lot cheaper than losing 4 in 4, and these numbers will only improve.

Does anyone know the relative cost between relaunching a pre-launched rocket and the total launch costs? Are reusing engines/rockets something like $10k/$50M, or $10M/$50M?

The initial min cost is 62 million. In the short term they are trying to reduce the price by 30%. That would make the price something like 40 million. The Upperstage is not reused and it costs about 10 million. 30 million for launch and refitting of equpiment seems to be reasonable.

Longer term they hope to reduce the price to much less.

Nobody really knows. Refueling is more like the first number, but it may need repairs, refits or otherwise.

Refueling, according to Elon, is about 200k.

Looks like they just need a bigger (wider) ship. Problem solved :)

No first stage has missed the drone ship so far. They were always right on target. So I don't think it needs to be bigger.

Reply on news.ycombinator.com