SpaceX Raptor engine - General Thread 4

[OTV Booster]

A left field thought: is there a mechanism that could add electrons to the O2 flow to stave off chemical reactions?

I'm told that methane does a fabulous job of giving electrons to O2.

Probably why they use it as rocket fuel.  lots of it too.

all joking aside, that's precisely what those finicky expensive materials around any hot GOX do - they develop a surface oxidized layer that protects the material.

(stainless does this in normal atmosphere, but hot GOX is another level of difficulty)

Ok, so it's not a yes/no type of thing. It's that good ol' trade space. My question morphs into: what combinations of materials and temps would work? This has the knock on question of: How much cooling can be realistically expected from a secondary layer of channels on the MCC?


It's totally beyond me but maybe some interested engineer has some ideas.

[OTV Booster]

Far be it from me to fault anybody's armchair quarterbacking, but SpaceX had a chance to fix the LOX tank icing problem on v3 and they didn't do it.  That should probably tell you something.

I doubt they made the wrong decision.  I'm very interested in why it turns out to be the right decision, but coming up with progressively more byzantine ways that a heat exchanger could work probably doesn't shed much light on that.

My best guess:  The added heat exchanger dry mass per engine exceeds the ice mass dumped into the LOX tank per engine.

Something got past me. How do we know this?

[LouScheffer]

It would be moderately insane to use LOX to cool anything important.

Perhaps not completely insane.  Check out the paper Liquid oxygen cooling of hydrocarbon fueled rocket thrust chambers.

[lightleviathan]

It would be moderately insane to use LOX to cool anything important.

Perhaps not completely insane.  Check out the paper Liquid oxygen cooling of hydrocarbon fueled rocket thrust chambers.

Launcher's E2 engine uses LOx cooling and has more difficult cooling constraints compared to Raptor. It's not completely insane, just requires smart engineering

[catdlr]

https://twitter.com/SpaceRhin0/status/2009626987208511856

Rhin0
@SpaceRhin0
Yesterday we saw another 6 Raptor 2's heading away from McGregor
These appear to have been already prepped for flight before they were put in storage.

[catdlr]

https://twitter.com/mcrs987/status/2010817654978732296

TheSpaceEngineer
@mcrs987
We're running Raptors on the Venezuelan oil now I guess

[TheRadicalModerate]

Far be it from me to fault anybody's armchair quarterbacking, but SpaceX had a chance to fix the LOX tank icing problem on v3 and they didn't do it.  That should probably tell you something.

I doubt they made the wrong decision.  I'm very interested in why it turns out to be the right decision, but coming up with progressively more byzantine ways that a heat exchanger could work probably doesn't shed much light on that.

My best guess:  The added heat exchanger dry mass per engine exceeds the ice mass dumped into the LOX tank per engine.

Something got past me. How do we know this?

I thought I had an authoritative source for this, but I can't find it.  Best I can find was the statements from Musk in the Everyday Astronaut interview from about 18 months ago:

Something I think we'll do in the future is move to, for critical valves, serious parallel valves.  So any one valve failure does not, no matter what happens, does not take out the ability of the ship to orient itself correctly.

If they were going to go from preburner tap-off to heat exchanger, I'd have thought he would have just said that.  There's also further conversation about the difficulty of producing autogenous GOX.  Again, this would have been a fine place to say, "We're fixing this in Raptor v3."  He didn't.  That's not definitive proof, but it's suggestive.

[catdlr]

Pete Hegseft's visit to Starbase today, and Space displays a Raptor power head.

https://twitter.com/mcrs987/status/2010865949927195078

TheSpaceEngineer
@mcrs987
shielding design changes between rsn20 & these flight engines, prev images 1&2, now images 3&4. slimmed down. no other visible design changes yet

[catdlr]

https://twitter.com/SawyerMerritt/status/2010877640337441049

Sawyer Merritt
@SawyerMerritt
·
SpaceX showed off today its next-generation Raptor 3 rocket engine, the most advanced rocket engine ever made. It will first fly aboard Starship Version 3, with the first launch targeted for Q1 2026.

• Almost 2x the thrust of Raptor 1
• Costs 4x less
• Much lighter. Will save 2,425 lbs of weight per engine, or 94,575 lbs (42.9 metric tons) per launch
• No heat shield
• Optimized for manufacturability

[SpaceLizard]

Some parts look blue? Is that a lighting artifact or a new material/coating?

[russianhalo117]

Some parts look blue? Is that a lighting artifact or a new material/coating?

That is stage lighting intentionally be used to add to the look.

[catdlr]

Some parts look blue? Is that a lighting artifact or a new material/coating?

Lighting

[TheRadicalModerate]

I couldn't goad anybody into taking the bait on this one up-thread, so I'm going to try again:

To avoid sooting, the fuel-rich preburner likely burns GCH4 and LOX at near-stoichiometric mixture (4:1), then mixes additional GCH4 downstream to cool the gas so it doesn't melt the preburner walls or the turbine blades.

I'm wondering if the LOX preburner might do something similar to generate autogenous O2 for pressurant, with as little icing as possible.  It doesn't have to, because the fuel-lean mixture will naturally burn to completion with no sooting, but if they do this, then there might be an intermediate, kinda half-mixed state between the stoich-burned combustion products and the pure injected O2.  That could potentially increase the tapped-off O:F mixture from 50:1ish to something like 100-200:1.  At that mixture, the amount of combustion products going into the tanks would be much smaller.

The question is whether there's a practical intermediate state:  Hot enough that the LOX is flashed to vapor and has enough enthalpy to support the needed tank pressurization, but not yet mixed completely with the hot CO2 and steam.  This is clearly not an equilibrium state, so you'd need gobs of CFD to figure out whether such a preburner design would be possible.

[OTV Booster]

I couldn't goad anybody into taking the bait on this one up-thread, so I'm going to try again:

To avoid sooting, the fuel-rich preburner likely burns GCH4 and LOX at near-stoichiometric mixture (4:1), then mixes additional GCH4 downstream to cool the gas so it doesn't melt the preburner walls or the turbine blades.

I'm wondering if the LOX preburner might do something similar to generate autogenous O2 for pressurant, with as little icing as possible.  It doesn't have to, because the fuel-lean mixture will naturally burn to completion with no sooting, but if they do this, then there might be an intermediate, kinda half-mixed state between the stoich-burned combustion products and the pure injected O2.  That could potentially increase the tapped-off O:F mixture from 50:1ish to something like 100-200:1.  At that mixture, the amount of combustion products going into the tanks would be much smaller.

The question is whether there's a practical intermediate state:  Hot enough that the LOX is flashed to vapor and has enough enthalpy to support the needed tank pressurization, but not yet mixed completely with the hot CO2 and steam.  This is clearly not an equilibrium state, so you'd need gobs of CFD to figure out whether such a preburner design would be possible.

Oooh. That would be ticklish.


How about this. Let the combustion products flash to vapor and pass them through a cold trap before they reenter the main O2 stream. The O2 would stay vapor, water & CO2 drop out as solids. It's a heat exchanger with additional duties and a novel place in the process. IDK if there is any mass or complexity savings, but just about anything should be better than the full tank filter they've been using.


Now if there were only some way to use the waste products to cool the heatshield on reentry without busting the mass budget or getting stupid with the plumbing.

[TheRadicalModerate]

How about this. Let the combustion products flash to vapor and pass them through a cold trap before they reenter the main O2 stream. The O2 would stay vapor, water & CO2 drop out as solids. It's a heat exchanger with additional duties and a novel place in the process. IDK if there is any mass or complexity savings, but just about anything should be better than the full tank filter they've been using.

You'd have to dump most of the gas enthalpy to pull the ices out.

What's a "full tank filter"?  I don't see why you wouldn't just filter the sumps and the inlets for the vents / cold-gas thruster valves.  That shouldn't be very much mass.

PS:  Water ice will float on LOX.  CO2 ice will sink.  Does that help with filtration?  Does it change how you'd condition the prop with ullage acclerations?

[rsdavis9]

Ok, I maybe missing something.

What is the volume, mass, percent of LOX needed for autogen?
1/1000th of total volume?
How big of a tube is needed for that?
It seems that putting this tube almost anywhere(MCC, Throat, Preburner) will get the needed heat so that when it expands from supercritical to gas it will be warm enough. 270K to 500K?

So ratio of gas(oxygen) at STP to liquid is 1/798. Assumedly it might be even more at possibly 500K. Twice as much.
1/1000 of 900kg/s is kg/s or 60kg/min so bigger than a garden hose. A garden hose can fill a 5gal bucket in a min? Which is 19kg/min.



1.429g/L oxygen at STP
1.141g/cm^3 at b.p.
https://en.wikipedia.org/wiki/Oxygen

EDIT: Yea I forgot the target tank pressure is 6bar. It's still in the ballpark!

[TheRadicalModerate]

Ok, I maybe missing something.

I assert we're all missing something, because that's not what SpaceX chose to do, at least twice, likely three times.  They thought that preburner tap-off and icing vs. a pump tap-off with a heat exchanger was a good trade.

Why?


BTW:

1m³ of LOX weighs 1141kg.  (Update:  at boiling, it's 1143kg/m³.  Close enough.)

1m³ of O2 at 6bar, 250K is 600,000Pa * 1m³ = n * 8.314J/k·mol * 250K, so n = 289mol.  O2 has a molecular weight of 32g/mol, so the mass is 9.2kg.

Liquid/ullage gas ratio is 124:1.

If we arm-wave the Raptor mass flow at 600kg/s, then we need 4.8kg/s of LOX for autogenous pressurant.

Enthalpy of vaporization for O2 is about 213kJ/kg, so we need 1.03MW of heating power through our hypothetical heat exchanger.  That's not nuthin'.

[Genial Precis]

Maybe I've missed something but it seems like the solution which has actually been implemented for the icing problem is the lox header tank. IIRC the biggest problems with icing in testing thus far were attributed by CSI Starbase to engine relight. So if they have a lox header tank isn't used until relight, that tank won't contain the ice generated by the main burn of the booster and so the icing problem is greatly mitigated, which is probably good enough for now.

It doesn't have to be true that that's their ideal solution. Maybe eventually they will decide to use one or more heat exchangers for ullage. Or not.

[OTV Booster]

Ok, I maybe missing something.

I assert we're all missing something, because that's not what SpaceX chose to do, at least twice, likely three times.  They thought that preburner tap-off and icing vs. a pump tap-off with a heat exchanger was a good trade.

Why?


BTW:

1m³ of LOX weighs 1141kg.  (Update:  at boiling, it's 1143kg/m³.  Close enough.)

1m³ of O2 at 6bar, 250K is 600,000Pa * 1m³ = n * 8.314J/k·mol * 250K, so n = 289mol.  O2 has a molecular weight of 32g/mol, so the mass is 9.2kg.

Liquid/ullage gas ratio is 124:1.

If we arm-wave the Raptor mass flow at 600kg/s, then we need 4.8kg/s of LOX for autogenous pressurant.

Enthalpy of vaporization for O2 is about 213kJ/kg, so we need 1.03MW of heating power through our hypothetical heat exchanger.  That's not nuthin'.

Just out of curiosity, what mass of propellant at the expected ratio would it take to produce 1MW of thermal? Not suggesting anything. Just trying to get a sense of scale.


My hope is that the decision to do combustion tapoff was a 'get it working for now' thing and not a long term solution. Ice in the LOX just feels so wrong, especially with a rapid reusability goal.

[LouScheffer]

[...]
If we arm-wave the Raptor mass flow at 600kg/s, then we need 4.8kg/s of LOX for autogenous pressurant.

Enthalpy of vaporization for O2 is about 213kJ/kg, so we need 1.03MW of heating power through our hypothetical heat exchanger.  That's not nuthin'.

It's not nuthin', but it's not much on the scale of rocket engines.  The combustion chamber and nozzle are already a giant heat exchanger.  One estimate uses a mass flow of 500 kg/s of methane absorbing heat as it rises from -161°C to +600°C, resulting in roughly 845 MW of heat absorbed by the fuel.  A measly MW is way down in the noise - in my guess, the drawbacks would be in complexity of plumbing and materials that can withstand oxygen.  The energy is there for the taking.