Booster 18 Testing Anomaly - Massey - Nov 21, 2025

[woods170]

This LOX landing tank eliminates the possibility of ICE floating on the fuel

Ice floating on the fuel? There is no ice in the methane tank

Fair point about correct usage of the word "fuel" vs "propellant" or "oxidizer". LOx technically isn't "fuel"

For clarity:
- Liquid Methane is the fuel
- Liquid Oxygen is the oxidizer
- "propellants" is the term used to point to all chemical components involved (in this example liquid Methane and liquid Oxygen) in the continueing combustion process inside the Raptor engines.

So, liquid Methane is a propellant, as well as the fuel, but definitely not an oxidizer. Liquid Oxygen is also a propellant, as well as an oxidizer, but definitely not a fuel.

[Slarty1080]

This LOX landing tank eliminates the possibility of ICE floating on the fuel

Ice floating on the fuel? There is no ice in the methane tank

Fair point about correct usage of the word "fuel" vs "propellant" or "oxidizer". LOx technically isn't "fuel"

So the LOX landing tank prevents any possibility of ice entering the engines because the LOX landing tank is more or less isolated from the main LOX tank. Although the main LOX tank gets contaminated with ices during the ascent, the landing tank remains ice free until the landing burn starts. At that point the landing tank will need gas feed to maintain the correct pressure as the landing tank LOX level drops. Either there's then insufficient time left for ice formation to cause a problem during landing or they still use screens and filters inside LOX landing tank to keep the ice out or they use COPV gas to maintain pressure during landing.

[OTV Booster]

This LOX landing tank eliminates the possibility of ICE floating on the fuel

Ice floating on the fuel? There is no ice in the methane tank

Fair point about correct usage of the word "fuel" vs "propellant" or "oxidizer". LOx technically isn't "fuel"

So the LOX landing tank prevents any possibility of ice entering the engines because the LOX landing tank is more or less isolated from the main LOX tank. Although the main LOX tank gets contaminated with ices during the ascent, the landing tank remains ice free until the landing burn starts. At that point the landing tank will need gas feed to maintain the correct pressure as the landing tank LOX level drops. Either there's then insufficient time left for ice formation to cause a problem during landing or they still use screens and filters inside LOX landing tank to keep the ice out or they use COPV gas to maintain pressure during landing.

Maybe a vent between the main tank and the top of the landing tank.


When the landing burn starts autog gas goes to the main tank as usual and as the landing tank level drops the vent at the top  allows entry of well pressurized main tank ullage. CO2 ice still snows down to the main tank sump and water ice still plates out onto the main tank dome, walls and sump fluid surface.


The contaminants entering the landing tank equalization vent would be minuscule except in the event of a big slosh. Even that would be a one time event and if the max height of the sump slosh is kept below the vent height on the landing tank, the impact would be negligible.


Simple filters in the landing tank outlet might suffice.

[Brigantine]

Maybe a vent between the main tank and the top of the landing tank.

I like this idea.

You're effectively using the residual LOx in the main tank as a filter to remove water vapour, then re-heating the H₂O/CO₂-depleted gas (by mixing in new hot autogenous gas) to pressurize the landing tank.

Add pipes to take the already-H₂O/CO₂-depleted ullage gas from near the top of the ullage space if slosh is a concern.
Still need to do some maths re water vapour mass fraction vs density. And we don't know how much the ullage gas cools down or what fraction of water vapour / CO₂ becomes floating/sinking ice.



OTOH, I'm not sure how much this is even necessary. The landing tank is a much smaller volume and has much less time for the ullage gas and LOx to interact - it might not be enough to clog simple filters in the landing tank, even with direct autogenous pressurization. It might be more mass-efficient to let the main tank pressure drop as it cools, since it isn't carrying much load for landing.

Let the landing tank be a pressure vessel inside a pressure vessel. No carbon overwrap necessary.

[InterestedEngineer]

Let the landing tank be a pressure vessel inside a pressure vessel. No carbon overwrap necessary.

is the landing tank new to Booster 18?  or was it on the prior models?

[rsdavis9]

Let the landing tank be a pressure vessel inside a pressure vessel. No carbon overwrap necessary.

is the landing tank new to Booster 18?  or was it on the prior models?

This particular design is brand new with the tank off to one side of the downcomer. I think in v2 it was at the bottom of the downcomer more or less coaxial. It had copv's mounted to the outside of the landing tank.

[InterestedEngineer]

Let the landing tank be a pressure vessel inside a pressure vessel. No carbon overwrap necessary.

is the landing tank new to Booster 18?  or was it on the prior models?

This particular design is brand new with the tank off to one side of the downcomer. I think in v2 it was at the bottom of the downcomer more or less coaxial. It had copv's mounted to the outside of the landing tank.

That's what I thought.

I think "a pressure vessel inside a pressure vessel"  rolls off the tongue far too casually.

It may be the root cause of Booster 18's demise.   Coordinating pressures across pressure vessels inside pressure vessels must be perfect.  Get it wrong, and a tank gets crushed.   (e.g. see the prior downcomer mishaps).  And there's all sorts of ways to get it wrong - internal software, stuck valves, external (stage 0) software, etc.

[OTV Booster]

Maybe a vent between the main tank and the top of the landing tank.

I like this idea.

You're effectively using the residual LOx in the main tank as a filter to remove water vapour, then re-heating the H₂O/CO₂-depleted gas (by mixing in new hot autogenous gas) to pressurize the landing tank.

Add pipes to take the already-H₂O/CO₂-depleted ullage gas from near the top of the ullage space if slosh is a concern.
Still need to do some maths re water vapour mass fraction vs density. And we don't know how much the ullage gas cools down or what fraction of water vapour / CO₂ becomes floating/sinking ice.



OTOH, I'm not sure how much this is even necessary. The landing tank is a much smaller volume and has much less time for the ullage gas and LOx to interact - it might not be enough to clog simple filters in the landing tank, even with direct autogenous pressurization. It might be more mass-efficient to let the main tank pressure drop as it cools, since it isn't carrying much load for landing.

Let the landing tank be a pressure vessel inside a pressure vessel. No carbon overwrap necessary.

The main tank is not so much a filter as a cold trap.


As per InterestedEngineer, a pressure differential between main and landing tanks introduces another point of failure. Maybe reduce pressure in both to the minimum necessary for proper engine feed. IIRC John Livingston, a true rocket engineer, figured Raptor 1 needed ~3bar inlet pressure at startup. Anybody have a clue what the g load is when the landing burn starts?


Needless to say, the equalization vent needs to be large enough to maintain adequate flow.

[rsdavis9]

b17 vs b18 from whataboutit video

edit: if you watch the video Felix says thats the methane header in the b17 pic so maybe there was no separate landing tank for LOX?

[rsdavis9]

As per InterestedEngineer, a pressure differential between main and landing tanks introduces another point of failure. Maybe reduce pressure in both to the minimum necessary for proper engine feed. IIRC John Livingston, a true rocket engineer, figured Raptor 1 needed ~3bar inlet pressure at startup. Anybody have a clue what the g load is when the landing burn starts?


Needless to say, the equalization vent needs to be large enough to maintain adequate flow.

Well at start it is de-accelerating at the max from at altitude atmosphere so probably at least 2g. So say 2-3g as a guess.
 

[OTV Booster]

Adding to the above post:


Venting the main tank from, for example, 6bar to 3bar, at engine start drops the main tank temperature which amps up its cold trapping. It also drops engine inlet pressure at exactly the moment g loading is fortuitously increasing. BUT, as the booster eases into the chopsticks the g load is decreasing as the fluid level in the landing tank is reaching minimum. Bummer.


Maybe the gods are willing to smile and decree that once started, R3 can operate at a very low throttle and inlet pressure. Excuse me if I decline to hold my breath.

[catdlr]

https://twitter.com/AshleyKillip/status/1997204350935175368

ChromeKiwi
@AshleyKillip

Booster 18's forward section with the new custom load spreader attached to the new integrated hot stage ring and what is remaining of the aft section rolled from Massey's test site to back to production. Great shots between the booster transport stand from
@StarshipGazer

The very Top and bottom are the only remaining items being returned to the production site.

https://x.com/CeaserG33/status/1997212732005781992

Ceaser G
@CeaserG33

Booster 18's aft and forward sections rolled from Massey's towards the Production Site tonight.

https://twitter.com/ENNEPS/status/1997234504789082599

Elisar Priel
@ENNEPS

A somber night at Starbase, as the remains of B18 took the "green mile" from Masseys towards Sanchez for final scrapping following its testing accident.

[Slarty1080]

Adding to the above post:


Venting the main tank from, for example, 6bar to 3bar, at engine start drops the main tank temperature which amps up its cold trapping. It also drops engine inlet pressure at exactly the moment g loading is fortuitously increasing. BUT, as the booster eases into the chopsticks the g load is decreasing as the fluid level in the landing tank is reaching minimum. Bummer.


Maybe the gods are willing to smile and decree that once started, R3 can operate at a very low throttle and inlet pressure. Excuse me if I decline to hold my breath.

Its an interesting idea, but I fear they might not be able to make use of the main LOX tank after boost back due to ullage collapse. As soon as the Raptors stop the autogenous gas stops and the temperature and pressure start to fall. I'm not sure how fast they will fall, but there will be a jolt when the engines cut after the boost back burn, followed by a flip and 3-4 minutes of free fall with grid fin adjustment / jiggling. Not sure what the pressure in the tank will be after that.

Maybe just use COPV gas to pressurize the landing tank until they can come up with something better?

[OTV Booster]

Adding to the above post:


Venting the main tank from, for example, 6bar to 3bar, at engine start drops the main tank temperature which amps up its cold trapping. It also drops engine inlet pressure at exactly the moment g loading is fortuitously increasing. BUT, as the booster eases into the chopsticks the g load is decreasing as the fluid level in the landing tank is reaching minimum. Bummer.


Maybe the gods are willing to smile and decree that once started, R3 can operate at a very low throttle and inlet pressure. Excuse me if I decline to hold my breath.

Its an interesting idea, but I fear they might not be able to make use of the main LOX tank after boost back due to ullage collapse. As soon as the Raptors stop the autogenous gas stops and the temperature and pressure start to fall. I'm not sure how fast they will fall, but there will be a jolt when the engines cut after the boost back burn, followed by a flip and 3-4 minutes of free fall with grid fin adjustment / jiggling. Not sure what the pressure in the tank will be after that.

Maybe just use COPV gas to pressurize the landing tank until they can come up with something better?

So, if main and landing tanks don't share ullage the landing tanks have to be pressure vessels that might be higher or lower pressure than the mains, and take on the combustion products - if that's still an issue. Ugh.

I don't think collapse will be a problem. No data, just a gut thing. Unless something radical happens the grid fins will be tweaking more than hammering and the flip won't have the urgency of hot staging where collapse is skillfully avoided. There will be a jolt at shutdown but I expect it to be in line with the long axis resulting in no or little slosh. Lastly, the engine first free fall should generate enough drag to keep things settled down.

It's all opinion and conjecture. YMMV.

[Slarty1080]

Adding to the above post:


Venting the main tank from, for example, 6bar to 3bar, at engine start drops the main tank temperature which amps up its cold trapping. It also drops engine inlet pressure at exactly the moment g loading is fortuitously increasing. BUT, as the booster eases into the chopsticks the g load is decreasing as the fluid level in the landing tank is reaching minimum. Bummer.


Maybe the gods are willing to smile and decree that once started, R3 can operate at a very low throttle and inlet pressure. Excuse me if I decline to hold my breath.

Its an interesting idea, but I fear they might not be able to make use of the main LOX tank after boost back due to ullage collapse. As soon as the Raptors stop the autogenous gas stops and the temperature and pressure start to fall. I'm not sure how fast they will fall, but there will be a jolt when the engines cut after the boost back burn, followed by a flip and 3-4 minutes of free fall with grid fin adjustment / jiggling. Not sure what the pressure in the tank will be after that.

Maybe just use COPV gas to pressurize the landing tank until they can come up with something better?

So, if main and landing tanks don't share ullage the landing tanks have to be pressure vessels that might be higher or lower pressure than the mains, and take on the combustion products - if that's still an issue. Ugh.

I don't think collapse will be a problem. No data, just a gut thing. Unless something radical happens the grid fins will be tweaking more than hammering and the flip won't have the urgency of hot staging where collapse is skillfully avoided. There will be a jolt at shutdown but I expect it to be in line with the long axis resulting in no or little slosh. Lastly, the engine first free fall should generate enough drag to keep things settled down.

It's all opinion and conjecture. YMMV.

I thought I would try asking ChatGPT to see what it makes of it (just for a laugh):
Question
“The SpaceX Superheavy booster propellent tanks are believed to run at an operational pressure of somewhere between 3 and 6 bar. When Superheavy finishes its boost back burn and the engines cut out the autogenous hot gas pressurization will stop leading to a fall in temperature and pressure in the tanks during the descent. The tanks will contain residual cryogenic liquid, will undergo a jolt during engine shut down, a flip manoeuvre and fall unpowered under gravity for 3-4 minutes. Is it likely that the tank pressurization will remain above 2 bar at this point?”

Answer
Very long answer <snip>
summary
“Thus, unless SpaceX has additional pressurant injected (e.g., COPV helium or an active heating/pressurization scheme after boostback — which is not publicly documented), tank pressures could easily fall below 2 bar during the 3–4-minute coast. Whether it stays slightly above atmospheric or falls below depends on the exact residual liquid volume and heat transfer, but maintaining 2 bar in that interval is not assured solely from residual cryogenic boil-off.”

LOL – I was moderately impressed that it had some sort of handle on the question, but was a bit underwhelmed (what did I expect! doh). So who can say at the moment? Maybe you are right but if I were a betting man…   

[OTV Booster]

Adding to the above post:


Venting the main tank from, for example, 6bar to 3bar, at engine start drops the main tank temperature which amps up its cold trapping. It also drops engine inlet pressure at exactly the moment g loading is fortuitously increasing. BUT, as the booster eases into the chopsticks the g load is decreasing as the fluid level in the landing tank is reaching minimum. Bummer.


Maybe the gods are willing to smile and decree that once started, R3 can operate at a very low throttle and inlet pressure. Excuse me if I decline to hold my breath.

Its an interesting idea, but I fear they might not be able to make use of the main LOX tank after boost back due to ullage collapse. As soon as the Raptors stop the autogenous gas stops and the temperature and pressure start to fall. I'm not sure how fast they will fall, but there will be a jolt when the engines cut after the boost back burn, followed by a flip and 3-4 minutes of free fall with grid fin adjustment / jiggling. Not sure what the pressure in the tank will be after that.

Maybe just use COPV gas to pressurize the landing tank until they can come up with something better?

So, if main and landing tanks don't share ullage the landing tanks have to be pressure vessels that might be higher or lower pressure than the mains, and take on the combustion products - if that's still an issue. Ugh.

I don't think collapse will be a problem. No data, just a gut thing. Unless something radical happens the grid fins will be tweaking more than hammering and the flip won't have the urgency of hot staging where collapse is skillfully avoided. There will be a jolt at shutdown but I expect it to be in line with the long axis resulting in no or little slosh. Lastly, the engine first free fall should generate enough drag to keep things settled down.

It's all opinion and conjecture. YMMV.

I thought I would try asking ChatGPT to see what it makes of it (just for a laugh):
Question
“The SpaceX Superheavy booster propellent tanks are believed to run at an operational pressure of somewhere between 3 and 6 bar. When Superheavy finishes its boost back burn and the engines cut out the autogenous hot gas pressurization will stop leading to a fall in temperature and pressure in the tanks during the descent. The tanks will contain residual cryogenic liquid, will undergo a jolt during engine shut down, a flip manoeuvre and fall unpowered under gravity for 3-4 minutes. Is it likely that the tank pressurization will remain above 2 bar at this point?”

Answer
Very long answer <snip>
summary
“Thus, unless SpaceX has additional pressurant injected (e.g., COPV helium or an active heating/pressurization scheme after boostback — which is not publicly documented), tank pressures could easily fall below 2 bar during the 3–4-minute coast. Whether it stays slightly above atmospheric or falls below depends on the exact residual liquid volume and heat transfer, but maintaining 2 bar in that interval is not assured solely from residual cryogenic boil-off.”

LOL – I was moderately impressed that it had some sort of handle on the question, but was a bit underwhelmed (what did I expect! doh). So who can say at the moment? Maybe you are right but if I were a betting man…   

LoL. I think you got an artificial maybe.


I'm a fan of feeding regen into COPVs and bleeding it off (maybe heated) for ullage. That would take care of ullage contraction but probably not collapse. It also gives a high pressure source for hot and cold gas thrusters, but that's more important for the ship than the booster.

[OTV Booster]

Was gonna ask this in booster/ship preflight prep thread but it fits better here.


Has COPV mishandling been pinned down for booster 18 or is it still speculation? Has there been any sign of stand alone COPV pressure testing?

[meekGee]

Was gonna ask this in booster/ship preflight prep thread but it fits better here.


Has COPV mishandling been pinned down for booster 18 or is it still speculation? Has there been any sign of stand alone COPV pressure testing?

I didn't aee anything to indicate whether the COPV failed under working conditions, he COPV was accidentally taken beyond its limits, it was a COPV-attached hardware that failed, or it wasn't even COPV related.

[catdlr]

https://twitter.com/Blobifie/status/2001904835315273819

Blobifi
@Blobifie
Over the past few weeks/days, workers have been seemingly replacing the COPVs on Ship 39 as we've seen multiple COPVs be lifted and lowered from around Ship 39's payload bay.

It is speculated that this may be a result of B18's incident at masseys, or it could be something else.

[OTV Booster]

Blobifi
@Blobifie
Over the past few weeks/days, workers have been seemingly replacing the COPVs on Ship 39 as we've seen multiple COPVs be lifted and lowered from around Ship 39's payload bay.

It is speculated that this may be a result of B18's incident at masseys, or it could be something else.

Musta heard me ask.


Not solid proof that COPVs are involved but it is a bit of a hint. I've got a picture in my mind of a hard hat guy squinched up next to a blast wall, fingers in ears, and a COPV on the other side in a cage, pressure line attached. And a big stack of COPVs waiting to test.