Hacker News Reader

Comment by runesoerensen | Hacker News Reader

"That core booster approached the platform as planned, but it unfortunately hit the water going 300 MPH and was lost, because some of its return engines failed to light"

https://techcrunch.com/2018/02/06/spacex-landed-two-of-its-t...

View on HN

Thanks for the closure - I've been refreshing twitter all morning.

I've seen the much-touted $90M price-tag for the FH launch, but does that take into account loss of the core or boosters?

I think he said in the press conference that they weren't planning to reuse the core or boosters. (But then mentioned later they wanted some of the parts? IDK)

Correct, the engines were going to be retired anyway but they are keen to reuse the titanium grid fins on the boosters which cost a small fortune to make.

The core actually had older ablative grid fins made of aluminium so no great loss there[0].

[0]https://i.redd.it/l6s2bds7pl801.png

Have you seen an approximate price for those fins anywhere. I saw the interviews, but I want to know how expensive is expensive.

Platinum is USD$32K per KG, so if the platinum grid fins are 100KG, then they would cost > $300K each just for materials. With 4 per booster that might be a million dollar price tag for the grid fins.

The fins are titanium, not platinum. The cost of material is negligible.

Oops, you’re right my mistake

That would be $3200k for each, but they are made out of titanium, not platinum, and most of the cost is in the work, not material. I'd like to know the price too though.

Yep - Titanium is notoriously difficult to work with.

Titanium isn't particularly terrible to machine. Nothing is hard to machine if your budget can cover the correct tools and you machines can provide the right setting to run those tools.

Your typical mechanical engineer fresh out of school has incredibly limited experience when it comes to things that are not stupid-proof to work with (engineering programs have other priorities). They then go on to build specialized knowledge in various subjects and usually more on the design side, not the execution side. Of course someone who designs plastic molds or simulated impeller designs all day is going to create a black box around things that aren't their specialty. You don't care about how the impeller or mold is made other than knowing that it can be made, what its material properties are and knowing that actually making it involves a bunch of details you don't know so you offload it to a 3rd parts (for the same reasons someone else is having you design the impeller or the mold).

A bunch of engineers and otherwise smart people on the internet saying that titanium is like computer programmers saying residential electrical is complicated or web devs complaining about bash. It really doesn't mean much but people who have no experience with this things tend to think the people who only have a shred know what they're talking about.

It's not difficult. Most other people just know they don't know how to do it and that they don't know what they'd need to know to go about learning how.

It's not that difficult. The more common alloys machine relatively easily; some of the tougher ones can be a pain but not much more than other special purpose materials like Inconel.

So difficult that the SR-71 leaked fuel before starts and had to be re-fueled right after take-off. https://en.wikipedia.org/wiki/Lockheed_SR-71_Blackbird#Fuel

The SR-71 took off with a small fuel load to reduce stress and improve engine-out performance, that is why it was refuelled immediately after take-off.

The mental image of it leaking so much fuel on the ground such that it needed immediate refuelling is a myth propagated extensively on the internet.

The characterization is/was also propagated by people who actually refuelled them, so I'm somewhat unconvinced it is a myth.

I believe if you look for primary sources like pilots and ground handlers/mechanics you'll mostly find reference to small leaks (I think they were referred to as drips caught in cookie trays).

I had a quick search and found the KC135 chap who says the refuelling was needed due to leaks, but without being rude to him I'm not sure he's really a qualified source for that information. It sounds more to me like 2+2=5.

I'm procrastinating so lets do some napkin maths, the claim is

1) a significant amount of fuel is leaking out of the expansion gaps,

2) climbing up to 25,000 ft at 300 knots would heat the airframe enough to seal those gaps,

3) there would still be sufficient expansion room to allow for travelling at M3.2

----

Ignoring that no engineer would be happy with 1.

For 2...Total Air Temp = Static Air Temp + Ram Rise. At 25,000 ft the static air temperature is about -35 C. Ram rise for a true airspeed of 300 kts:

RamRise = V^2 / 87^2 = 300^2 / 87^2 = 12 degrees.

So skin temperature at typical refuelling altitude would be -23 C

Titanium has an expansion coefficient of 9E-69 meters per meter-kelvin. So approximating rather grossly, assuming a s tarting temperature of 20 C over the 33 m length of the plane there would be a contraction of about 1 cm.

----

And for 3:

The aircraft then accelerates up to 1900 knots. RamRise = 1900^2 / 87^2 = 470 degrees

Static air temp at over 60000 ft is roughly -55 C. So skin temperature would be 420 C. So assuming the same length and starting temp, the plane would expand by around 11 cm

----

So to summarise: According to the claims, at ground level and temperature the expansion gaps were large enough to significantly leak fuel. After take-off the aircraft needs to be refuelled immediately. Assuming this is done so (i.e. take-off, climb to 20,000, refuel) then the skin temperature is lower than ground level, and the expansion gaps should have grown ever so slightly. The aircraft then climbs up to its M3.2 cruise point and everything expands significantly "as designed" and the gaps disappear.

----

Perhaps the anecdote we'll see on the internet now is that the SR-71 had to take off and go supersonic to rapidly heat up the skin before briefly decelerating to refuel, but the refuel had to be done super fast to stop the skin cooling down too far...

My primary source is my father, who was a senior NCO responsible for fueling operations. It was his job to know.

Fair enough, no point trying to dispute a family anecdote!

The SR-71 leaked fuel before takeoff because it needed separation distance between parts to allow for expansion. It had nothing to do with the difficulty machining titanium.

From your source: "Fuselage panels were manufactured to fit only loosely on the ground. Proper alignment was achieved as the airframe heated up and expanded several inches.[30] Because of this, and the lack of a fuel-sealing system that could handle the airframe's expansion at extreme temperatures, the aircraft leaked JP-7 fuel on the ground prior to takeoff."

The Blackbirds leaked because they had engineered gaps to allow for heat expansion.

I bet SpaceX could engineer something in spec that wouldn't leak fuel or need to be immediately refueled though.

And we’re also comparing an aircraft designed in the late 1950’s and built in the early 60’s by slide rule (while still being extremely accurate) to modern spacecraft having the benefit of decades of materials and design technology

That has nothing to do with titanium and everything to do with material expansion, the design/shape of the assembly and ambient vs operating temp.

Cant reply directly to wand3r for some reason. Re his comment spacex could engineer something that wouldn't leak or need instant refuelling - I wouldn't be surprised if they could - but they have around 30 years of advancements to help them!

They could do it back then too.

If you dig past the internet comments and read some of the "primary source" books, the picture of the leaks is very different. I can't remember the exact book I read it in, but the author states there was a tank sealant, and it lasted around 50 hours (I think), before it needed to be replaced.

This is somewhat backed up by the Jenkins book [1] which talks about the time consuming process of replacing sealant, and the Graham book [0] that is the source for the Wikipedia claims on expansion. It talks of different sealants used, and how leaks were precisely noted and collected in _shallow_ drip trays.

[0] https://books.google.co.uk/books?id=dX5cCgAAQBAJ&pg=PA129&dq...

[1] https://books.google.co.uk/books?id=lEpWE748QUsC&q=sr71+seal...

KU Leuven states the grid fins at 41kg. (The same volume of material would be 193kg in platinum)

ahh, I noticed that in the feed and was wondering why. I guess they expected it was the least likely to work? or needed the better grid fins less because of the cylindrical top

Can they still recover them?

If I understood properly none of the boosters were block 5 so they weren't looking at re-flying them.

They are however interested in recovering the grid fins on the side boosters, which were redesigned to accommodate for the nose cones now sitting on top of them.

Smart that they used old pre-block5 parts for this test flight. (Not that I expected any less from those folks!)

Block 5?

The latest hardware revision that's expected to be the mainstay for F9 flights going forward. Has better performance and addresses some reusability issues discovered in earlier blocks.

But if they weren't on this flight, how sure is SpaceX that they are more reliable?

But yeah, that's just a question from my mind. This launch was amazing, the future in the making right here.

As far as I know, the center core was a block five one, but they didn't intend to refly it anyway, from what Elon said in the press conference. Maybe they intended to study it? If that's the case, there will probably be another "expandable" block five flight.

According to reddit, even the center core (B1033) was Block 3. No block 5 cores have flown yet.

https://www.reddit.com/r/spacex/wiki/cores#wiki_.5Bfh.5D_b10...

The plain is to lock down block 5's design so it can be man rated which in itself would make it more reliable. This will probably not happen until they fly them a few times though so that wouldn't mean the first block 5s would be more reliable.

I recall reading or hearing yesterday that space-x no longer plans to man-rate any falcon-9s instead reserving that for the BFR.

I believe that was in reference to the Falcon Heavy, not the F9. The F9 is expected to be man-rated for NASA's commercial crew program.

https://en.wikipedia.org/wiki/Falcon_9_Full_Thrust#Block_5

Version 5 of the falcon design.

To be perhaps slightly more accurate, it's like a major version 5.

Almost (if not) all missions include changes of some kind; the manufacturing blocks each have significant, incompatible differences in parts.

Upgrading the grid fins would be a minor version change, modifying the engine to increase its throttle depth a major one. I don't know, but it seems like there would need to be significant re-tooling for each new block, and many parts would be incompatible with previous blocks.

Even if they're not reusing any of it (this time), it's still very useful to examine and analyze how well it's held up.

Perhaps some piece of metal is showing more fatigue than they'd like to see, or a wire had damage to the insulation from a vibration. I'm sure that's the sort of thing they'll be looking at.

The $90m price is for a fully reusable FH, yes. The cores in this particular flight, however, were not intended for reuse either way.

SpaceX has been keeping their launch prices pretty consistent, even when they couldn't know how many F9 stages they would actually recover. I think they've factored in testing losses into the price.

I wouldn't be surprised to learn that advertised prices are still far above what they could achieve after flying for a while and paying down some of their R&D debt (and if/when the market becomes even a little bit competitive).

Almost inevitably. They don't need to hit rock-bottom prices because they're already there compared to the rest of the current market, and they can use the money to pay back their enormous R&D budget, or if Musk is just throwing that money off as a worthwhile loss then to pay for research on BFR and their other future projects.

After all if you can make a hefty profit while still undercutting the competition you're basically winning at capitalism.

The damage to the ASDS might be more of a problem. Depending on how badly it was damaged, it might not be available to catch other cores that could have landed otherwise

It look cool to have them land on their own, but wouldn't it be much cheaper to use parachutes and drop them in the ocean !? Either way they are going to take them apart down to each bolt and then reassemble again to make sure there are no faults. You don't simply fill it up and relaunch.

> You don't simply fill it up and relaunch.

Yes, you do. Or at least, that's what they want to do. They've already reflown a bunch of their landed rockets, including the two boosters on Falcon Heavy. The ultimate goal is to be able to fill it up and relaunch, and the ultimate motivation for propulsive landing is simple - that's what is needed on Mars (parachutes won't do much there), so they want to master it.

EDIT: I meant that today, with Falcon Heavy Test Flight behind us, the two side boosters qualify as reflown - this was their second mission, not the third.

So the FH boosters were reused? The ones that landed back again? Amazing, great news to start a morning :)

Yes!

The left booster "originally launched on July 18, 2016 in support of the CRS-9 mission, and landed back at LZ-1". The right booster "originally launched on May 27, 2016 in support of the Thaicom 8 mission. Notably, it earned the nickname "Leaning Tower of Thaicom"; having developed a significant lean upon a hard first landing."

Source:

https://www.reddit.com/r/spacex/comments/7vg63x/rspacex_falc... linking to details of:

- https://www.reddit.com/r/spacex/wiki/cores#wiki_b1025 - left booster

- https://www.reddit.com/r/spacex/wiki/cores#wiki_b1023 - right booster

No.

This has been gone ove many times, but:

Parachutes aren't cheap or easy. Saltwater is horrible for precision parts. They are trying to gas and go, or at least move, gas and full power test.

SpaceX goal with the Falcon 9 is to be able to land, simply check the core (the engine gets more complicated checkups), and to have it back on the launchpad within of a few days or weeks.

Arianespace (a major european launch company, known for the Ariane 5) has developed their own new engine concept for reusability, which is designed to be restarted unlimited times. Once Prometheus flies, they’ll not only have a reusable rocket comparable to the Falcon 9, but one that can fly, land, be refueled without checkup, and immediately fly again.

SpaceX also aims for that in the long run, but not in the Falcon 9 series itself.

> Once Prometheus flies

Sounds like they are not that committed to the project and if it happens at all it will be a long way off. Will it be worth competing with both SpaceX and Blue Origin?

“We could replace Vulcain 2.1 by Prometheus,” Bonguet told SpaceNews. “Or Prometheus can be the first brick to build the next generation. We will see where we are in 2025 or 2030, and then decide on the right time whether to go one way or the other.”

http://spacenews.com/ariane-6-could-use-reusable-prometheus-...

Prometheus is finished, production on the prototype has started, and the first flights of the prototype are expected in 2020.

If Arianespace follows that schedule, they’ll be a decade late compared to SpaceX, but still ahead of all other competitors in this.

There's no flight scheduled. They're developing a full-scale article to be ground-tested in 2020.

Prometheus or not, Ariane 6 is an expendable rocket in any form that currently or will soon exist. By the time they figure out the very basics of first-stage reusability on a launcher that is roughly equivalent to Falcon 9 FT, SpaceX will have BFR ... likely for quite a while.

At least their reusability plans are not a joke like ULA.

AFAIK, there's no major disassembly between flights. Certainly inspection and test firing, but nowhere near the meticulous detail and overhauls that the Space Shuttle went through after each flight. This is a completely different beast than what many observers' mental models are accustomed to.

A large part of the work SpaceX are trying to do is get it so that they can pretty much just give it a fairly simple inspection, fill it up and relaunch it. They already need far less work between launches than the space shuttle did.

I'm kind of glad they had something fail. That means maybe they learned something new to improve the next shot.

The rocket ran out of engine igniter fluid. It had enough to light the centre engine, but not the two side engines.

That seems to imply they were trying to do a 3-engine landing, which they tested recently in the 'failed to expend the rocket' incident. This might explain why they miscalculated the igniter requirement since they don't have much experience with 3-engine landings.

On the other hand, if you look carefully at the landing video, it looks like they did a three-engine landing boost on the other boosters: https://youtu.be/BBA7su98v3Y?t=517 So they have definitely learned something valuable in that landing incident.

Interesting. Looks like a 1-3-1 burn (start with 1, ignite two more for a bit, then finish on a single one).

I wonder how much longer the '3 engine burn' was for the GovSat launch last week that did a 'water landing'

Cool. I'd assumed they only did that with the centre core, but you're right.

Is the normal landing a 1-engine landing? If so, is the advantage of a 3-engine landing that it requires less fuel due to having a much shorter 'suicide burn'?

Yes, and yes. The less time you're in the air, the less fuel needed to counter gravity's continual pull, so the most efficient landing burn is basically the highest acceleration you can handle without coming apart such that your velocity hits 0 as your altitude does.

What’s a “suicide burn”?

most of the people I know are twisted pessimists. you're a twisted optimist

'Crazy things can come true': Elon Musk discusses Falcon Heavy launch: Full presser

https://www.youtube.com/watch?v=sytrrdOPYzA

Was it just lucky that the core didn’t directly impact the drone ship or was it designed to aim to the side in the event of the landing speed being too high?

It always comes in slightly to the side of the barge and as part of the landing burn it does a final sideways correction. I believe it was for this kind of situation - or one where the final burn doesn't even happen.

I read somewhere that the drone ship was damaged.

Elon mentioned in the press conference that two of the engines on the droneship (it has four, one on each corner) were damaged.

Do they send another ship to fix the drone ship?

There's another ship nearby which tows it, it doesn't go all the way in and out of port on its own power.

And the ship has taken far bigger poundings in the past, one of them blew a huge hole in the deck.

Oh ok, cool. I guess I thought it was more autonomous than that by the name.

Watching the other boosters land was really impressive.

I think it's just fuel saving for the ship rather than lack of autonomy

More time than fuel I think. The droneship has precision-manoeuvring thrusters, not engines that would let it move at any significant speed.

> But the core, middle booster, which attempted to land aboard “Of Course I Still Love You,” a drone barge that SpaceX uses as a mobile, ocean-borne landing pad stationed in the Atlantic for its flights departing from Florida, wasn’t recovered.

“Of Course I Still Love You” is the name of a spacecraft in The Culture, a series of novels that inspired Elon Musk when he was young.

Reply on news.ycombinator.com