Business, Space, Technology and Society

So Much For SN4

52 Comments

It went kablooie in the static test today. Video anon.

[Update a few minutes later]

It appears to be an earth-shattering kaboom.

Not sure why there’s no audio, but I assume it was loud. Presumably no one was hurt; Mary was reportedly quite a ways away when she took this. Starhopper looks OK, but we’ll hear more soon. I was watching the live stream; it seems to be still burning. One of the benefits of stainless is that it’s cheap, compared to carbon composite.

[Update a few minutes later]

Here’s before and after.

[Update a while later]

Loren Grush has the story.

[Update a few minutes later]

And here’s Eric Berger’s story.

[Update a couple minutes later]

Worth noting, as Eric does, that the static test was successful, FWIW. So this is the biggest advance they’ve made yet on the SN series. Next challenge is to do a static test without an explosion afterward.

[Saturday-morning update]

Here’s one with audio. Note the delay due to the distance. Has to be over half a mile away.

[Update a few minutes later]

An article about the safety of the system, with a quote from Leroy Chiao: “Chiao said pushing too hard for safer numbers could cause a spaceflight program to spiral into never launching at all.”

You don’t say.

52 thoughts on “So Much For SN4”

  1. Looks like, several seconds post-engine-burn, a high-volume liquid methane leak from the external stand plumbing, with the plume eventually igniting from the flare-vent downwind of the stand, and the detonation travelling back up the plume.

    Methane-air will detonate over a fairly wide mix range.

    One reason not to flare vented methane – you’re providing an ignition source for any leaks. My impression is that over the years flaring vents may have caused more explosions than they prevented.

    1. So I’m presuming the flare is for tank venting and burnoff at a safe distance rather than near the rocket.

      OTOH had that flare not been there how long does it take a cloud like than to no longer be explosive? You don’t want that just drifting around in the wind looking for a random ignition source.

      1. SN4 made a good Daisy Cutter. Things to remember about high vapor phase rocketry.

      1. I remember this mama of a BLEVE. Liquefied Petroleum Gas instead of methane, nothing like heavier than air disbursement and fire. Town on / town off….

        https://www.youtube.com/watch?v=viFC-RtzlFo

        My old man got a picture of 1/2 a tank car protruding from someone’s living room…. As the announcer says: “-> another <- small mid-western town." You know, like get used to it, they pop off all the time….

    3. Yes. Those who would like to see us transition to a hydrogen based energy system (the fact that it is energy transport upside down aside), ought to be aware of this little table:

      https://www.engineeringtoolbox.com/explosive-concentration-limits-d_423.html

      Methane is explosive from 4.4% to 16.6 % mix. (see results video) as opposed to Propane from 2.1% to 10.4% (see results video), compared to Hydrogen which scores two entries: 4% to 75% and 6% to 40%. Yeah, let’s pipe hydrogen around everywhere…. At least one can hope it will accumulate in your upstairs bedroom and kill you there in your sleep rather than just maim you when your house collapses around you.

      Also a shout out for best-in-class to Hydrazine (5% to 100%) and Propyl Nitrate (2% to 100%).

  2. “Perhaps another test protocol failure?”

    Which is worse, a stupid mistake that shouldn’t have happened or a serious design flaw?

    1. I am suspecting a random failure of a component feeding oxygen from the tank to the engine. The worst of all worlds.

    2. The worst is running away from the unknown unknowns.
      Don’t give in to the temptation of the unknowable. Knowing is our original sin and our destiny.

  3. Look at that flare stack in the second video. They’re venting oxygen, wind carries it to flare stack, fire spreads from flare stack to rocket.

    1. According to Henry Vanderbilt what I’m seeing is venting methane, not oxygen. Same idea though.

  4. I’d go with methane. Oxygen would have just made the flare burn hotter and brighter, there’s nothing around there to oxidize. Methane cloud drifting and mixing with oxygen from venting and air, then getting ignited by the flare.

  5. I have to ask…

    Does this one count as a setback? Or do we all do the “you learn more from failure than success” dance?

    2. It’s a setback when it’s an anomaly after years of practice. Means you overlooked something you should have known about. This is all learning curve.

  6. Slow motion video and frame-by-frame analysis from other angles seems to indicate that the fire started under the rocket, not over at the flare stack. That something hot would be underneath a rocket that had just conducted a test-burn is not surprising.

    In any even, I assume SpaceX has all kinds of camera feeds on the pad, looking up at the engine, so they probably know exactly what happened.

    1. Detonations can fool you on zoomed-in standard-rate vid. They move so fast they’ll happen across the entire field of view from one frame to the next. You have to look for secondary clues like the direction debris gets propelled.

      But there is a long shot available that shows the detonation shockwave expanding. Look carefully the initial detonation on the long-distance vid at https://twitter.com/KXAN_News/status/1266470457873698817?s=09

      The epicenter of the expanding shockwave hemisphere is clearly to the right of the vehicle, over the flare vent, not over the vehicle itself.

      The detonation started at the flare vent.

      1. Henry, that does indeed look as if the initial locus was offset from SN4.

        You’ve changed my mind on that – I was initially of the opinion that the locus was under the vehicle.

        Thank you.

      2. Oops, I was very positive about being wrong here. See correction farther down. (And excuse my y-axis dyslexia putting in an appearance, as I seem to have subbed “right to left” for the correct “left to right” in it…

  8. I’m a big fan of Musk, but maybe Space X should slow down a little (like taking one month instead of three weeks), and consider more carefully what they’re doing.

    2. If they’re not failing with test articles, then they’re not pushing the envelope hard enough.

  9. Ya know, if you really insist on having a flare stack, you should have four of them with selector valves, and only use the one *upwind* of the test stand.

    Just sayin’.

    At XCOR we abhorred the use of flare stacks, we had a nice loooong vent line off to the north (the wind is hardly ever from the north at Mojave) and only ever put liquid methane or hydrogen into the vent when we were doing a dump at the end of the day. Never once a cryogen fire.

    1. But on stepping through the videos frame by frame, it does seem that the explosion was centered below the vehicle, and the flare stack was still burning at that moment undisturbed by the blast. So it may well be that the fuel-air cloud was ignited near the Raptor and not the flare.

      1. Which vid do you see this in, Doug? Pointer appreciated.

        That said, standard-rate vid can be deceptive for understanding extremely fast events. Things can appear to happen first in one place on the screen just because of when that was electronically scanned relative to the rest of the image. I’m still inclined to trust more the visible blast-front hemisphere being centered over the downwind flare stack, not over the rocket, in that ~2km long shot.

          1. Also from that angle the curvature of the shock front does not appear to center on the flare.

          2. Thanks for the pointer!

            My, they do go on, indeed. Enthusiastic, yes. Knowledgable, not so much. (A rule for watching things like this: Ignore the commentary – believe the images. And even there be careful where the commentators are selecting the images.)

            But, something they said did provide a vital clue: Youtube has simple keyboard single-step controls. Didn’t know that. Pause the vid, then hit “,” for one frame back and “.” for one frame forward.

            I’m going to go play with that on the raw vids for a while now and see what I can see. Short version, from what I’ve seen on a quick pass, it does seem a more complex event than I thought. (Possibly more complex than they think too.) May take a while. More when I think I know more.

          3. OK. After single-stepping through the main two different vids they have at that nasaspaceflight.com youtube, yes, it looks like ignition first happened at the base of the vehicle. Either inside the vehicle base, or (more likely I think) from the GSE connections to the vehicle base. This was a relatively slow deflagration, at least at first – it’s four 30fps frames later before the weight at the top of the vehicle, still visible above the (relatively) slowly expanding fireball, definitely moves.

            Meanwhile, one frame after the first ignition glow is visible at bottom left of the vehicle, there’s a visible detonation front travelling down the mixed-air-and-methane leak plume, away from the vehicle, moving right-to-left. This looks to be what causes the first part of the upward-expanding blast wave visible to the right of the vehicle.

            The upward-expanding blast wave then continues, somewhat stronger, from around the vehicle itself – possibly as the initial deflagration and/or the plume detonation disrupt the vehicle structure and release/mix more fuel and LOX, which then also to some extent goes higher-order.

            So, yeah, always be careful of people’s first impressions right after these things 🙂

    2. Doug,
      I hope you’re well. Thank you for sharing. Your nuggets of rocket engineering knowledge are priceless, and you’re one of the few people that contribute to positively to the signal-to-noise ratio of the dumpster fire that is the aRocket* mailing list.

      *Henry Vanderbilt also gets a shout out for quality posts! But mostly aRocket is the Democrat Socialist Obama co-conspirator Bill Claybaugh just telling people they’re stupid and wrong without adding any lessons or context. Also there is some troll who keeps telling amateurs to do liquid hydrogen. If you’re a rocket hobbyist, aRocket is *not* the place to hang out, despite the list name.

      1. Tsk. Kind words, but there’s a lot more useful and interesting people contributing to Arocket than that. Even Bill C (he does have axes to grind, but not generally the ones you ascribe) is good to debate with – smart guy, keeps you on your toes.

        As for the perpetual hydrogen enthusiast, well, yes, it is the internet and there is inevitably a certain amount of noise to tune out…

        1. I enjoy Ken Mason’s stories, but a lot of time they’re build up with no punchline – I really wanted hear what happened on the MIPCC test but he just kept teasing us. I appreciate Henry Spencer’s history lessons.

          But nobody is really working on any projects that they share. Bob Watzlavick is still making progress. Bill C only does solids (because of his political connections) and even so he doesn’t what to share or hear what other people think. When he asks for input it’s rehetorical only – he just wants everybody to think he’s the smartest person in the room.

          I’m almost loathe to say it…but I miss Monroe’s presence on the list.

    3. “Ya know, if you really insist on having a flare stack, you should have four of them with selector valves, and only use the one *upwind* of the test stand.”

      Upwind? Why so? Maybe I don’t understand the use of the flare stack.

      1. The vent stack is to vent off fuel vapor a safe distance from your test stand. The flare is to then immediately burn the fuel at the stack, rather than risk a fuel-air mix drifting with the wind and igniting somewhere else less convenient.

        And if the flare stack is upwind of your stand, if your stand ever has a large fuel leak, the fuel plume will blow away from the flare, rather than toward it where it can ignite and flash back up the plume.

        I’m the paranoid type – I also envision problems where you vent unexpectedly large amounts through the stack and it drifts back toward your stand. I’d say, always use a stack that’s perpendicular to the wind (which a northerly stack generally is in Mojave) so neither stack nor stand venting drifts toward the other point. But then I’ve always had a healthier sense of self-preservation than Doug

    4. Doug at XCOR did you have a contingency plan to detonate a methane cloud or was the plan to allow it to dissipate? How long would that take and over what distance? It will be pushed away from the high pressure leakage source then dissipate upwards.

      2. No, no plans to deliberately set off a fuel-air explosive. We did that accidentally once at RotRock (silane mixed with GN2 and air can be flammable but not pyrophoric) and it was much too exciting. Put the vent well the hell off from anything that can ignite it, and “the solution to pollution is dilution” will let the lighter-than-air fuel vaporize and mix with air to get below the LEL.

        We briefly had a vent port on the methane test stand, but later plumbed it into the pipe away after we saw the shimmer of mixing cold CH4 into the air dip down to the ground while it was denser than air. A long pipe away and a tall vent puts all that mixing far away and high up to keep it clear of any ignition sources.

        1. While we’er at it, I can’t help but dream up Rube Goldberg workarounds to massive leak fails. Why not a drop canvas, that rings the rocket if a massive leak occurs? Forces the venting to go up rather than sideways. If its strictly a methane leak it would quickly flood the base with concentrations too high to ignite and quench ignition sources with cold gas. Simultaneously venting the lighter than air methane up the top. This is for non-crewed test vehicles. Also presumes no simultaneous LOX tank failure. Hm or maybe a silo instead of an open pad?

  11. If I were to retire to Boca Chica, I’d be seriously tempted to open a high end hamburger restaurant there. I’d call it Rudfuggers….

  12. Sound delay was ~10 seconds, so more like 2 miles away.

    And still loud enough and strong enough to set off car alarms; wild.

  13. Put that blast in a tunnel and you can see why coal miners are leery of methane. 6% makes best booms. High concentrations cant ignite. 3% gets you panicky. You would not put that naked flame near a gassy mine. Bet todays 2 travellers wish they hadn’t seen that.

  14. A couple of days ago I was arguing that surely they wouldn’t just put a 22 ton piece of steel on top of a tank, as that introduces complex deformations and stresses. I figured that if they were going to add lots of mass, they’d add it around the perimeter where its weight would transfer more directly to the cylindrical portion of the tank, similar to how an interstage keeps the loads on the outer skin. Or they could weld up a ribbed structure that would resemble a large roof structure, which would transfer a central load out to the walls. The could’ve added that to the mass before they placed it on top.

    I even cited a 1965 NASA report called The Stability of Shallow Hemispherical Shells Under Concentrated Load. I figured the extra weight of a support framework shouldn’t be an issue because the whole point was to add weight. But I was proven wrong because they stuck the mass right on top, kind of like a hat on an Easter Island statue.

    Then, under load, the mass stayed still while the top of the tank moved upwards, adding significantly to the pyrotechnic display.

    Interestingly, it seems the mass went up and then came down and smashed down right on top of the Raptor engine. ^_^

    1. Did you notice the jet that pushed the mass assembly nearly straight up from the top? It ignites shortly thereafter so Scott Manley called it a methane jet. Impressive. Not my recommended LES however.

  15. “Chiao said pushing too hard for safer numbers could cause a spaceflight program to spiral into never launching at all.” Does anyone really believe these numbers, or imagine that “pushing harder” could achieve safer numbers? Most fatal problems are things you didn’t anticipate. The way you learn about them is through repeated trials, hopefully not fatal. You are never going to learn till you start trying.

Comments are closed.