31 thoughts on “Starship”

  1. MIG welding in the wind is especially problematic because it keeps blowing the shielding gas away from the arc and puddle. Stick welding is generally a better option outdoors, but then you have slag to deal with, plus the base of the rocket would be surrounded by not only cigarette butts, but thousands of little welding rod stubs.

    Given their construction method, it probably doesn’t matter whether they did the welding from the outside or the inside, so maybe they should throw a tarp over the top and come up with an internal crane or scaffold.

    My approach would be somewhat different. It would build the rocket from the bottom up, just as they are doing, but I’d do it with the rocket upside-down. Start with the engine mounts (upside down) and then mount and pipe the engines to the base of the rocket. Then keep lifting the stage (with engines pointed toward the sky) and welding on new sections. The welders would be working at ground level and that bottom area would be surrounded by clamps, jigs, fixtures, and test equipment. Only when a section is done would the stage be raised up a level. All the slag and trash would just fall on the ground where it could be swept up, instead of falling into the base of the propellant tanks where all the pipes head towards the engines.

    In WW-II we built destroyer escorts upside-down, using a big ring assembly to roll them right-side up when almost finished. The idea there was that since all the complex piping on a ship was overhead, which is a difficult and uncomfortable position for supporting and welding, it would be vastly easier if all that piping and ducting was laying on the floor where it was easy to assemble. Building them upside-down turned out to be wildly faster than right-side up.

    1. That’s how we do it now except that instead of rotating the whole ship, we build large prefab sections upside down, rotate them, and then connect them to the rest of the ship.

  2. I wonder if SpaceX thought of calling these guys:


    The video of the steel tank rising from the ground is pretty amazing. With this tooling the Starship could be built from the top down, lifting up each built section as the next section is built below it. This would allow 80% to 90% of the welds to be made at ground level where the work is much easier.

    1. Sounds good for a mature design. But enlarging the structure to make a larger ship requires major investments in time and materials. Building from the bottom up allows major re-design (change of structural materials from carbon fiber to stainless) with only minor time and labor.
      I personally believe one of NASA’s problems is lack of flexibility to change direction when a flaw is discovered. This is caused partly by large, expensive, inflexible infrastructure which has only one use.
      For 50,000 dollars, Spacex can buy a taller lift if they need a taller rocket. If SLS needs to grow 50 feet taller, I don’t believe it would fit in the buildings NASA currently has at their disposal.
      I believe we have seen neither the final shape nor size of what will become the mature Starship.

    2. Years ago on the Selenian Boondocks Blog they had a post describing a method whereby tubes of any diameter could be created. The tubes would be constructed horizontally by continuously feeding in sheet metal in a spiral manner like a toilet paper roll. Any length of tube could be made by chopping the tube at the desired length. Continuous stir-friction welding would weld the sides of the sheet metal together thereby making the tubes. When the end of the sheet metal came then a separate step would be required to weld the next roll of sheet metal in an end-to-end manner.


      1. Elon has said that they’ll be switching from standard steel sheets to using whole rolls from the mill, and that the rolls would be a custom thickness. The mailing tube idea would eliminate all the vertical welds, so no weld would be perpendicular to the circumferential stress. It’s also easy to tweak the mill rollers to leave a slightly thicker edge, so that the stress on the weld is some smaller fraction of the stress throughout the sheet.

  3. NASA: Halt the program until the problem is identified in a multi year investigation and give the contractor a bonus to sit around and do nothing.

    SpaceX: We already started working on the replacement. This one was just a model to dork around with and we considered blowing it up anyway because it would be fun but the PR department wouldn’t let us.

  4. My guess, based on timing, is, they decided after the gas pressure leak tests mk1 would never be flight worthy. Note the next road closure was canceled, then they decided to pump in LN2 to see what would happen. I’m also guessing mk2 will be transported in pieces to LC-39A before final assembly and testing.

  5. I have some vague recollection from years ago that someone was going to build an orbital rocket using segmented stainless steel tanks made by a milk tanker truck company, but I can’t remember enough to do a keyword search into the archive. Anywone?

  6. I can’t yet verify this, but I’m hearing that the root cause was massive overpressure caused by an error in the pressurization equipment. The height the dome reached (400+ feet) sure as heck looks far in excess of what 6 bars (which I believe they intended to test to – double the 3 bar nominal) could do.

    *if* this failure was caused by the ground test equipment, it was still a bad day, but IMHO not as bad as if Mk1 had a structural failure at 6 bar.

    1. A universal throw model and its applications

      Excerpts from the abstract and introduction:

      A deterministic model has been developed that describes the throw of debris or fragments from a source with an arbitrary geometry and for arbitrary initial conditions.

      The model has been formulated for four basic source geometries: a point source, a vertical cylinder, a horizontal cylinder, and a vertical plane.

      The throw of debris and fragments is often the most
      dominant effect in explosion events. In case of deliberate
      explosions like those of ammunition shells, fragment throw
      determines the weapon effectiveness. In case of accidental
      explosions, like those in ammunition magazines and
      industrial explosions, the throw of debris plays an
      important role in the explosion hazard and consequently
      the required safety distances. As a result, the understanding
      and quantification of throw is of practical importance.
      The initial conditions for throw are defined by the
      distributions of the mass, launch velocity and launch
      direction. These conditions are determined by a broad
      range of complex physical phenomena. Examples are the
      fragmentation process in case of (stacks of) ammunition
      shells, the break-up process of concrete buildings, and
      vessel rupture which is typical for industrial explosions. In
      general, the failure process is determined by the explosion
      load on the structure in relation to its strength. This
      process results in distributions of mass and launch
      direction. The subsequent acceleration by the expanding
      reaction products largely determines the distribution of
      launch velocities. This is accompanied by pressure relief as
      a result of the increasing vent area between the accelerating

      We know about how high the top of the tank went, so we could probably get a good estimate of the internal pressure at rupture. By “we” I mean I’m not sure I’ve got time to get to it this evening. ^_^

      1. I am afraid I don’t have the time to do the math either (which in my case,(though not yours) actually means I don’t know how but don’t want to admit it). 🙂

        A factor that would weigh into this; drag effects on the dome, for which we’d need to know its mass as well as an approximation of its aerodynamics.

        There’s a thread over on NASAspaceflight trying to calculate the launch velocity, in order to estimate the pressure. I’m hopeful that they will come up with something.

  7. “Speculation: Bad welds from working in cold, dark, and wind.”

    “See, the problem with speculation is that it made a spec out of u and some guy named ‘lation.'”

    Dr. Gregory House, House

    1. OTRAG was much longer ago and a different design. What I’m remembering may have been a single article proposing to build an orbital LV from segmented, wedge-shaped stainless steel tanks fabricated by a milk-tank company in Minnesota. I don’t remember where I saw it or when (late 1980s?) and keyword searches bring up milk tanker trucks for sale and ice cream products with “rocket” in the name.

      1. Maybe you’re thinking of Sea Dragon. It wasn’t to be made by milk truck fabricators, but by shipbuilders. No launch pad, just tow it out to sea and fill the tanks, then light the candle under water.

        1. Again, I’m not looking for familiar steel rocket proposals of yesteryear, but a specific proposal for one with segemented, wedge shaped stainless steel tanks, and the milk tank manufacturer is about all I can remember. If I could remember where I saw it, I coud probably hunt it down. The old L5 Society magazine had a lot of oddball stuff in it, and I probably still have a collection of them. Somewhere, in a box, in a storeroom…

  8. Correction – OTRAG was intended to be made from steel, but the Wikipedia article doesn’t mention if it was stainless.

  9. Post CyberTRk, I wonder if the Starship hull has to be curved. Could it be facetted, with each facet sized to take a single heat shield tile on the windward side, fabricated from large sheets of stainless using the laser cutting and folding method to minimize welding?

  10. Was the loss of Starship a setback?

    No, it shows that they’re making rapid progress.

    There are two ways to tell if SpaceX is making rapid progress:

    1. Their tests succeed.

    2. Their tests fail.

      1. I would guess Starship; SLS may never fly.

        But your point escapes me. Is SLS your gold standard that all other systems should be compared to? If so, you picked a very low hurdle to clear.

        1. It’s the one that NASA, honestly or otherwise, says we should clear.

          But it is your contention that Ship/SH, if it meets its all stated goals, would not be a huge advancement over the past?

          Now, I’m struggling to understand your point.

          1. But it is your contention that Ship/SH, if it meets its all stated goals, would not be a huge advancement over the past?

            Of course it would; who said otherwise? I’m questioning your assertion that the loss of the Mk. 1 wasn’t a setback.

          2. If it was, Elon doesn’t seem to think so. He’s already had a better version under construction, and it’s not clear that flying Mark 1 was on the critical path to flying the improved versions. He said that while it would have been nice to fly Mark 1, it was primarily a pathfinder for construction techniques.

          3. If it was, Elon doesn’t seem to think so.

            I think the operative word here is “seem”.

            He’s already had a better version under construction, and it’s not clear that flying Mark 1 was on the critical path to flying the improved versions.

            I think the failed pressure test had some purpose.

            He said that while it would have been nice to fly Mark 1, it was primarily a pathfinder for construction techniques.

            I don’t think the failed pressure test can be construed as an unqualified vindication of those construction techniques.

            I find it hard to believe that Musk had no interest in how this test turned out, that success or failure was all the same to him.

  11. It depends on how you define “success” or “failure.” The purpose of testing is to learn, and the only true failure of a test is one from which you learn nothing. Generally, failure is a better teacher than success.

  12. It depends on how you define “success” or “failure.”

    Surely you’re not going to claim that the pressure test was a success?

    The purpose of testing is to learn, and the only true failure of a test is one from which you learn nothing.

    I suspect you would not be nearly so generous had this happened to Boeing, Lockheed-Martin, or NASA.

    Generally, failure is a better teacher than success.

    Sure, but it does not follow that failure is more desirable than success.

    1. I suspect you would not be nearly so generous had this happened to Boeing, Lockheed-Martin, or NASA.

      That’s a good guess, because all of those entities would have taken much more time, and spent much more money (and taxpayer dollars) to fail. And they probably wouldn’t have learned much.

Comments are closed.