Starship 36 explosion at Massey - June 18, 2025

[1]

Same question, though:  can you detect fatigue cracks via remote sensing of some kind?

Depends on how you define 'remote'.

Eddy-current testing is commonly used to detect shallow forming cracks, though you need access to both interior and exterior of the tank in question to do the job properly. But it would work; at least for a titanium tank. For things like COPVs, it becomes much harder to see fine or formnig cracks in general because the strands of the weave tend to set a limit on what your resolution is regardless of NDT technique. Surface wave ultrasonic inspection is a thing, but works much better on homogenous-ish materials. All of this requires access to the tanks though, and inspection in general gets more difficult logistically post-install.

If we define "remote" in a strict sense, like "scan is performed after the ship is buttoned up and otherwise ready for launch", then probably not. If "remote" is relaxed a bit, to something like "tanks have been installed, but technicians still have physical access to them", then probably yes. YMMV.

[TheRadicalModerate]

I talked to a guy from Luxfer last month, he said that "some of their customers in the space sector" were often asking if it would be safe to exceed the rated temperature limits.  I suspect that might have some bearing on this failure.

If you're heating cryogenic liquids, you can engineer things so that the heating remains isothermal until everything transitions to supercritical, can't you?  You'll still want the highest rated temperature, so that adiabatic expansion into the pressure lines doesn't drop the temperature low enough that it collapses back to liquid, but you can add heat at the same time, maintaining supercriticality.

[TheRadicalModerate]

Same question, though:  can you detect fatigue cracks via remote sensing of some kind?

Depends on how you define 'remote'.

Eddy-current testing is commonly used to detect shallow forming cracks, though you need access to both interior and exterior of the tank in question to do the job properly. But it would work; at least for a titanium tank. For things like COPVs, it becomes much harder to see fine or formnig cracks in general because the strands of the weave tend to set a limit on what your resolution is regardless of NDT technique. Surface wave ultrasonic inspection is a thing, but works much better on homogenous-ish materials. All of this requires access to the tanks though, and inspection in general gets more difficult logistically post-install.

If we define "remote" in a strict sense, like "scan is performed after the ship is buttoned up and otherwise ready for launch", then probably not. If "remote" is relaxed a bit, to something like "tanks have been installed, but technicians still have physical access to them", then probably yes. YMMV.

It's not just the time between flight pressure and launch that you have to worry about.  In both the depot and HLS versions, COPVs will have to provide pressurant for settling, RCS, and possibly landing thrusters.  The mass of pressurant in the COPV will then be a function of how long it stays on-orbit, which will be the Ship's entire lifetime.

At some point, you have to recharge the COPVs.  That's why I suspect that the ultimate pressurants are methane and oxygen:  flow 'em into the COPVs at low pressure, seal the COPVs, and heat until supercritical.  That cycle of multiple expansions and relaxations on the walls of the COPV is going to be something that needs useful telemetry.

The alternative is everything is pump-fed.  For RCS and settling thrusters, you can probably use electric pumps.  But for the landing thrusters, you may need gas turbines to generate the necessary work through the entire landing burn (~460kN for up to 30-ish seconds).  That's probably beyond the specific power of a reasonably-sized battery pack.

[1]

It's not just the time between flight pressure and launch that you have to worry about.  In both the depot and HLS versions, COPVs will have to provide pressurant for settling, RCS, and possibly landing thrusters.  The mass of pressurant in the COPV will then be a function of how long it stays on-orbit, which will be the Ship's entire lifetime.

At some point, you have to recharge the COPVs.  That's why I suspect that the ultimate pressurants are methane and oxygen:  flow 'em into the COPVs at low pressure, seal the COPVs, and heat until supercritical.  That cycle of multiple expansions and relaxations on the walls of the COPV is going to be something that needs useful telemetry.

The alternative is everything is pump-fed.  For RCS and settling thrusters, you can probably use electric pumps.  But for the landing thrusters, you may need gas turbines to generate the necessary work through the entire landing burn (~460kN for up to 30-ish seconds).  That's probably beyond the specific power of a reasonably-sized battery pack.

I understand. That's the main reason I want to know the use cases you have in mind, because it alters your options. And I have my answer. Your definition of 'remote' is about as strict as it can get: on orbit.

Let's ignore ship 36 for the moment, since that's almost certainly not an issue related to fatigue.

Tanker flights to refuel a depot will be coming back to Earth regularly, where they can be inspected as frequently as desired. Depot vehicles and similar will need more longevity on orbit (and may end up never returning to Earth at all). So what I would do in the beginning, at least until more heritage is established, is try to build both the depot and the tanker ships out of common part lots of COPVs, such that the latter can be used as a coal-mine canary for the former. Not just the same basic design; we want them built as similarly to one another as possible. In this, I mean use the same batch of composite wrap, same batch of epoxy matrix, cure/heat them in the same autoclave at the same time, etc, etc.

So if you can build, say, 5 COPVs to a lot, then put 1 of those COPVs in a depot ship and put the other 4 in tankers. Then, do periodic inspection on the tanker COPVs (after every flight, so desired), and if one of them starts to show signs of wear, then you at least have a heads up that it's time to retire and replace (or repair, if somehow possible) the corresponding depot ship. And not just COPVs. I would do the same thing to absolutely every part of every type that can be batched. The reliability folks will have a bit of work to do ironing out the details on that, but it might be better than trying to add a bunch of specilized telemetry to the depot ships.

All that said, I can also much more easily see them willing to add a crapload of extra telemetry to the HLS and crewed Mars-bound ships, so again, YMMV.

[Coastal Ron]

It's not just the time between flight pressure and launch that you have to worry about.  In both the depot and HLS versions, COPVs will have to provide pressurant for settling, RCS, and possibly landing thrusters.  The mass of pressurant in the COPV will then be a function of how long it stays on-orbit, which will be the Ship's entire lifetime.
...

I understand. That's the main reason I want to know the use cases you have in mind, because it alters your options. And I have my answer. Your definition of 'remote' is about as strict as it can get: on orbit.

Let's ignore ship 36 for the moment, since that's almost certainly not an issue related to fatigue.

Tanker flights to refuel a depot will be coming back to Earth regularly, where they can be inspected as frequently as desired. Depot vehicles and similar will need more longevity on orbit (and may end up never returning to Earth at all)...

For the ship you need COPVs that are inside the ship, since you can't have them hanging on the outside of the ship during launch and re-entry. But for depot versions, once in orbit why not attach an external COPV assembly that can be "easily" replaced as the COPV age?

This is definitely OT, but just pointing out that the restrictions that apply to a ship that launches and lands don't need to be the same restrictions that apply to what are essentially orbital stations.

[clongton]

Just replace the COPVs by building in the necessary volume at the top of the propellant tank behind a stainless steel common bulkhead. Sooner or later they are going to HAVE to do this anyway because the ultimate aim is for totally reusable spacecraft that NEVER return to earth where they can be inspected and replaced if found to be fatigued; spacecraft that will be refueled and depart for DOZENS, if not eventually HUNDREDS of trans lunar or interplanetary missions before being decommissioned due to old age. The COPVs are an Achilles Heel for long term reusability. Bite the bullet. Make the change - and never worry about a burst again - ever.

My 2 cents
YMMV

[KilroySmith]

I was reading a story about the OceanGate Titan sub implosion, and they discussed that the sub had microphones that could (and did) pick up the sounds of individual carbon fiber strands breaking in trips prior to the final, fatal one.
Could an audio or vibration sensor, operating continuously, detect a COPV in need of replacement?  It may not help in the case of a defective COPV that doesn’t meet its specifications, but it might catch initial sounds of failure during pressurization, allowing SpaceX to halt operations, stand down and replace.

[BN]

SpaceX Falcon 9 team is already quite experienced with COPVs and I think they do some kind of monitoring. There was some Turkmenistan (I think?) satellite mission that had some COPV that was "trending poorly" and they swapped it out prior to a test fire I think. They do maintenance and test them every 10 launches or so.

[BN]

It’s a question of factor of safety. A hydrogen car has COPVs and they aren’t inspected each fill up and can be refilled like dozens or hundreds of times in between inspection. Because the factor of safety is conservative enough. Flight applications are more mass sensitive, so they may choose greater inspection frequency to get by with less factor of safety. But it’s not inherent to COPVs that inspect has to be between every flight. In fact, pretty much all pressure vessels need inspection, not just COPVs.

refilled dozens of times between inspections??? those tanks are supposed to survive high-speed collisions on the highway.

you're supposed to be able to smash that tank between two delivery trucks going 200kph, otherwise how could they possibly be on the road.


here's a video of them shooting it with a 50cal.. which is not a scenario anyone is actually worried about.

[ELinder]

The COPV label photo shows a temp limit of only 149 F degrees? Is that always, or when at full pressure? I'm willing to bet a black COPV sitting in the summer south Texas sun for even a short time can get hotter than that.

[ulm_atms]

snip.....

The picture of the label from that video is from a damaged COPV.  "Made in the USA"

Manufacture sales sheet:  https://ngtnews.com/luxfer-offering-new-g-stor-high-pressure-hydrogen-cylinders#:~:text=Luxfer%20Gas%20Cylinders%20has%20introduced,350%20bar%20(5%2C000%20PSI).

The image displays a label for a Luxfer G-Stor® Go H2 cylinder, a Type 4 carbon composite cylinder designed for high-pressure hydrogen storage.
Key information from the label and search results:
Product: Luxfer G-Stor® Go H2 cylinder, Serial Number LRR1585.
Purpose: Certified, lightweight hydrogen storage solution for applications like fuel cell transit buses, heavy-duty trucks, vans, bulk gas transport, boats, and trains.
Specifications:
Volume: 728 L.
Empty Weight: 155.8 Kg.
Service Pressure: 380 Bar.
Liner Temperature Limits: -40 °C to 65 °C (-40 F to 149 F).
Safety Warning: "Do not fill if damage has caused strand unraveling".
Date Information: Manufactured in March 2025 and should not be used after March 2045.

Um...That's a small temp range for spaceflight applications...........  As Elinder said...Texas sun can easily get it hotter then that.  My front door handle can reach 190F in direct sun and it's dark brown.

And please tell me they aren't using this for cryo....anything...? 

[envy887]

And please tell me they aren't using this for cryo....anything...?

They are visible in-flight in the payload bay video stream on previous flights. They aren't insulated, and there's no ice or frost on them, so either the air in the payload bay is really dry, or they are close to ambient temperature.

[ulm_atms]

And please tell me they aren't using this for cryo....anything...?

They are visible in-flight in the payload bay video stream on previous flights. They aren't insulated, and there's no ice or frost on them, so either the air in the payload bay is really dry, or they are close to ambient temperature.

I only asked as there is many a talk of loading as cryo and letting it get warmer to reach correct gas pressure(I know they do this for F9's He...AMOS-6 anyone?).  I have ZERO clue as to how many types of COPVs and their ratings they have on Starship except this one though.  Just figure I would ask the question.

I'm more curious how they would keep them above -40 on the dark side of Earth.  Without any type of heater, it would get cold.  Having to use the contents of it would make it even colder.  -40 is not cold for Space.

[rsdavis9]

So with all this talk about COPV's what is the penalty for titanium or aluminum? I know aluminum scuba tanks are a lot better than steel.

[Robotbeat]

It’s a question of factor of safety. A hydrogen car has COPVs and they aren’t inspected each fill up and can be refilled like dozens or hundreds of times in between inspection. Because the factor of safety is conservative enough. Flight applications are more mass sensitive, so they may choose greater inspection frequency to get by with less factor of safety. But it’s not inherent to COPVs that inspect has to be between every flight. In fact, pretty much all pressure vessels need inspection, not just COPVs.

refilled dozens of times between inspections??? those tanks are supposed to survive high-speed collisions on the highway.

you're supposed to be able to smash that tank between two delivery trucks going 200kph, otherwise how could they possibly be on the road.


here's a video of them shooting it with a 50cal.. which is not a scenario anyone is actually worried about.

you’re making my point for me.

And what is your objection anyway? That i included “dozens” in the range?
the value I saw while researching the topic for my post was about 36,000 miles in between inspections, which is just under 100 fill ups if going from zero to 100%. So I couldn’t quite say “hundreds” by itself, and thus mentioned “dozens or hundreds” to make clear there’s a range of values.

[Robotbeat]

So with all this talk about COPV's what is the penalty for titanium or aluminum? I know aluminum scuba tanks are a lot better than steel.

Note that all those options require regular inspection.

[SpaceLizard]

It’s a question of factor of safety. A hydrogen car has COPVs and they aren’t inspected each fill up and can be refilled like dozens or hundreds of times in between inspection. Because the factor of safety is conservative enough. Flight applications are more mass sensitive, so they may choose greater inspection frequency to get by with less factor of safety. But it’s not inherent to COPVs that inspect has to be between every flight. In fact, pretty much all pressure vessels need inspection, not just COPVs.

refilled dozens of times between inspections??? those tanks are supposed to survive high-speed collisions on the highway.

you're supposed to be able to smash that tank between two delivery trucks going 200kph, otherwise how could they possibly be on the road.


here's a video of them shooting it with a 50cal.. which is not a scenario anyone is actually worried about.

you’re making my point for me.

And what is your objection anyway? That i included “dozens” in the range?
the value I saw while researching the topic for my post was about 36,000 miles in between inspections, which is just under 100 fill ups if going from zero to 100%. So I couldn’t quite say “hundreds” by itself, and thus mentioned “dozens or hundreds” to make clear there’s a range of values.

I don't think he was disagreeing with you... I think he believes you undersold their durability...

[BN]

It’s a question of factor of safety. A hydrogen car has COPVs and they aren’t inspected each fill up and can be refilled like dozens or hundreds of times in between inspection. Because the factor of safety is conservative enough. Flight applications are more mass sensitive, so they may choose greater inspection frequency to get by with less factor of safety. But it’s not inherent to COPVs that inspect has to be between every flight. In fact, pretty much all pressure vessels need inspection, not just COPVs.

refilled dozens of times between inspections??? those tanks are supposed to survive high-speed collisions on the highway.

you're supposed to be able to smash that tank between two delivery trucks going 200kph, otherwise how could they possibly be on the road.


here's a video of them shooting it with a 50cal.. which is not a scenario anyone is actually worried about.

you’re making my point for me.

And what is your objection anyway? That i included “dozens” in the range?
the value I saw while researching the topic for my post was about 36,000 miles in between inspections, which is just under 100 fill ups if going from zero to 100%. So I couldn’t quite say “hundreds” by itself, and thus mentioned “dozens or hundreds” to make clear there’s a range of values.

well, allow me to continue making your point if I may.

regardless of the number of refills without inspection, the COPVs used in hydrogen vehicles are intended to be able to withstand high speed collisions between vehicles, at pressure, without rupturing (which would kill everyone in both vehicles, plus some).

while I don't know how those COPVs compare to the specific ones used in Starship, it doesn't seem like COPVs themselves need to be eliminated from the design. that being said, a car crash involving hydrogen COPVs is probably my least preferred variety.

[Robotbeat]

The answer to this is simply that hydrogen car COPVs are much heavier for the same volume and pressure than the ones SpaceX was using (in the way they are using them). This goes for aerospace in general.

And to anticipate others: If you tried to design a rocket with universal adherence to ASME boiler code for pressure vessels (3.5-5 is the required factor of safety, 2.4 in limited circumstances), it literally wouldn’t leave the ground. There is not a single company or government who has done so. Atlas is around 1.25-1.4 iirc, and Falcon is 1.5. It also can change based on material.

[BN]

The answer to this is simply that hydrogen car COPVs are much heavier for the same volume and pressure than the ones SpaceX was using (in the way they are using them). This goes for aerospace in general.

actually toyota lobbied the japanese government to lower the safety standards for those tanks because they were too expensive/heavy. allegedly. but yes, probably stronger than the starship COPV. I guess your point that I am making here is that they can user stronger COPVs, and it seems very premature at this stage to be deleting them from the design for something much heavier than a stronger COPV.


unrelated to COPVs but relevant to the thread; there was another big rocket program that seemed to be on track at first, but then faced repeated setbacks for various reasons until it was unsustainable. although, if apollo hadn't already won they likely would have kept going for a few more launches.

these were very large explosions happening with a full stack, completely destroying the launch pad. (once)