Hacker News Reader

Comment by mabbo | Hacker News Reader

For those who aren't following this religiously like I am, here's the Tl;Dw:

SpaceX's falcon 9 is pretty big, but Starship is massively larger. It's also made cheaply out of steel in a huge garage in rural Texas. The engines are the most advanced ever made (arguably). Tomorrow, SpaceX is going to try to fly the prototype 12.5km up, then have it freefall, falling sideways. At the last moment, it will reorient itself and land on a landing pad.

Or, it will explode stupendously.

It's going to be awesome either way.

View on HN

Just to add a bit more color on size: Starship itself is about 50m in height, somewhat shorter than the full Falcon 9 stack (70m). Starship has a much larger diameter though, at 9m vs 3.7m for Falcon9.

But Starship is just the second stage. That's what's being tested tomorrow.

Eventually this will sit atop a massive new super-heavy booster. The full stack will be 122m high, nearly 2.5x taller than Falcon 9, and with a dramatically larger volume.

> It's going to be awesome either way.

That might be a better definition of science than “It’s like messing around, but taking notes.”

I love how everyone quotes it, and SpaceX deserves the credit for making that attitude popular. But what really makes it possible is cheap cameras — so when watching a blast that is likely destroying every living thing and destroying every piece of electronic in its way, I’ll be cheering for GoPro too.

I think the increased telemetry ("3000 channels" per the Falcon 9 AMOS-6 explosion; who knows what Starship will have) and the ease nowadays with which one could store the data on an SD card tucked inside every multiply-redundant avionics box, makes it much more likely you'll be able to pinpoint the cause of an error even after a catastrophic explosion. Hence enabling SpaceX to fail-fast with confidence that they will be able to find the problem from the data and fix it.

Plus apparently they're relatively cheap to build, that just makes the test even cheaper.

https://arstechnica.com/science/2016/09/spacex-with-a-manife...

If you care about telemetry then you can’t store it, you have to beam it back.

And SD cards are a nightmare from a reliability standpoint. Nobody (in this context) would do that.

https://www.vice.com/en/article/pga9dk/what-its-like-to-find...

They at least used SD cards in GoPro cameras on fairings. I admit I cannot authoritatively speak to what they are using for telemetry storage if they decided to go that way.

SD card connections are unreliable in an environment with heavy vibrations (something you’d expect with a rocket). For something like a GoPro, that might be acceptable to SpaceX, but for critical telemetry data, I highly doubt it as GP stated.

SpaceX is hiring software engineers to build the system that will ingest telemetry from Starship, and much more. Apply here https://boards.greenhouse.io/spacex/jobs/4943043002?gh_jid=4...

Also important to note that this is Starship Serial Number 8. If it has an "accident" it's good to know that there is actually a Starship SN9 that has just been completed at the manufacturing base a couple km from the launch site, standing by to take its place and try again.

Then there are two more vehicles abou 50% done - just join the tanks and add the pointy part in the front.

And parts have been spotted assigned up to SN16!

So this is really no ordinary space program where you prepare for years to launch a single super expensive rocket and then deliberate for months or years what went wrong and how to fix it.

No, they built the best rocket engine ever, attach it to glamurous 50s SF rocket built from stainless steel (and random Falcon 9/Dragon/Tesla parts) in a marsh and then launch and make more until it can go to space cheaply and reliably.

The world's most sophisticated engines on a vessel that started out as a water tower. I can't get enough of Starship.

> It's also made cheaply out of steel in a huge garage in rural Texas.

This makes it sound like it's made out of old barn roofs and recycled Quonset huts... which would be kind of awesome.

Well, the test launches have all looked like someone stuck a rocket on a grain silo...

A new or recently polished grain silo.

> The engines are the most advanced ever made (arguably).

Maybe not arguably. They embody a technology that's been considered for a long time the Holy Grail of rocket science - the full-flow staged combustion cycle:

https://en.wikipedia.org/wiki/Staged_combustion_cycle#Full-f...

As much as I love the Space Shuttle, I wish the US had not went all in on Hydrogen Oxygen. The RS-25 are amazing engines, but we could have had operational full flow staged engines 20 years ago if we had gone down that route. Though probably not the landing part.

Interesting that full flow LOX-LH2 engines could be easier to get, as oxygen and hydrogen can be ignited in a wide range of ratio. With LOX-kero you might have to have a two stage gas generator, when first you ignite components with a good ratio and then add a component to the already established combustion.

> With LOX-kero

Small nitpick: kerolox is much nicer sounding and what people know them as ;)

If you are going to nitpick the kerolox, why not hydrolox too? ;)

(I’m not one of the downvoters for the record)

Oh I just missed it :)

LH2 is way less dense than RP1 though, so you’d need a significantly larger fuel tank to carry the same amount of fuel.

Hydrolox (LOX-LH2) is about 30% more power than Kerolox (LOX-RP1), but is about 3.5x less density.

Source: https://en.m.wikipedia.org/wiki/Liquid_rocket_propellant#Bip...

Hydrolox - Ve: 3,816 d: 0.29

Kerolox - Ve: 2,941 d: 1.03

If you were wanting to optimize for high specific impulse, it seems LOX-LH2:Be or a fluorine based fuel is the way to go purely on the numbers (not accounting for cost or stability).

NASA should’ve kept producing the Saturn V. Over time, the production costs should’ve come down.

The shuttle was a rather mediocre trade off, in terms of LEO payload capacity.

Although granted, the shuttle landings were pretty cool.

Thanks! Just started reading about rocket engines, I don't think most people realize how many moving parts the SpaceX engines have!

Okay, I'll bite, How many moving parts do the SpaceX engines have?

Assuming the fuel and oxidizer impellers are on a common shaft, that's one. There must be a throttle valve somewhere, so that's two. What am I missing? Are there reduction gears?

They use pintle injectors in the Merlins, so that does move as well (though only a short distance) and is likely what makes them so throttleable.

https://en.m.wikipedia.org/wiki/Pintle_injector

Not sure why I was downvoted. Are the moving pintle injectors separate from a throttle vavle?

I have studied a lot of pintle injector designs and it can go either way. Reguardless, a moving pintle injector typically only has two moving parts.

TL;DR: Liquid-fuel rockets work by combining a liquid fuel with oxygen. To get the fuel and oxygen into the combustion chamber in huge quantities, they use a big fancy pump. To power this big fancy pump, all liquid-fuel rockets burn fuel in a separate combustion chamber, using the hot gas to power the pump. After the gas powers the pump, it has to go somewhere. Most rockets just dump the gas (these are called 'Gas-generator cycle' engines). This is nice and easy (ha), but it also means that the gas isn't going through the main combustion chamber, so it isn't propelling the rocket. The rockets on Starship re-use the gas that drives the pumps, pushing them into the combustion chamber and burning any unused fuel/oxygen to propel the rocket.

> The rockets on Starship re-use the gas that drives the pumps, pushing them into the combustion chamber

How is that possible? From a total layman's point of view, the pressure in the combustion chamber should always be greater than whatever the pump can put out, thus leading to a pump stall... what am I missing?

Actually the pump has to be able to put out as much if not more pressure than the pressure generated in the combustion chamber, otherwise it wouldn't be able to push a steady stream of fuel and oxidizer into the chamber.

I recommend watching or reading https://everydayastronaut.com/raptor-engine/

As the other post says, this cannot be true, otherwise how could the engine work. You'd not get any propellant into the combustion chamber.

> To power this big fancy pump, all liquid-fuel rockets burn fuel in a separate combustion chamber, using the hot gas to power the pump.

There's some engineers in New Zealand who might have something to say about that...

https://en.m.wikipedia.org/wiki/Electron_(rocket)

Fascinating! Their rocket's trust-weight ratio (~73, according to Wiki) is WAY below the state of the art gas-generator cycle Merlin (~185), presumably because of the added weight of the motor and batteries required to operate an electric turbopump. It's always interesting to see folks look past a supposedly bad design and discover that its other benefits (in this case, cost and simplified design, apparently?) can still render it a viable option.

Here is a picture of a Merlin engine on a test stand. To the right of the nozzle, the black and sooty exhaust of the turbopump can be seen.

https://old.reddit.com/r/engineteststands/comments/8f7c1q/me...

Everyday Astronaut did a nice explanation of the Raptor and various simpler designs, and what's so great about full-flow staged combustion:

https://everydayastronaut.com/raptor-engine/

This is the internet. I say 'arguably' because someone out there wants to argue.

And look! You found a friend!

That is not strictly true.

It’s a Heinlesque story.

Yes, indeed, we have plenty of real-life cyberpunk around us already, may it please be the time when earlier space SF finally starts to become real, too? :)

>Tomorrow, SpaceX is going to try to fly the prototype 12.5km up, then have it freefall, falling sideways. At the last moment, it will reorient itself and land on a landing pad.

>Or, it will explode stupendously.

We got both!

Or, you will not go to space today.

Reply on news.ycombinator.com