Sunday, July 18, 2021

Mass Production of Rocket Engines is No Longer An Oxymoron

SpaceX is planning to break ground soon on a new manufacturing facility for their Raptor engines in McGregor, Texas.  The plan is to produce 800 to 1000 Raptor engines in a year.  From that link to Elon Musk's Twitter account:

We are breaking ground soon on a second Raptor factory at SpaceX Texas test site. This will focus on volume production of Raptor 2, while California factory will make Raptor Vacuum & new, experimental designs.

Musk also referred to the coming factory as the “most advanced rocket engine factory in the world” and while it may sound like hyperbole, I find that very believable simply based on quantities.  Until now, rocket engines have been more like hand-crafted boutique items than something mass produced using the most advanced technologies.  I wouldn't be surprised if in their first year or two they produce more engines in the thrust class of the Raptor than have been produced in all of history. 

A fully equipped and fully modern factory is what you need if you intend to build and outfit a vast interplanetary fleet of Starships and the earthbound Super Heavy boosters that will send them on their way to Earth orbit, the Moon, Mars, and beyond.  It makes the Raptor family picture posted here on the 12th look like "an OK start." 

While the picture seems to show nine sea-level Raptors along with one vacuum version (lying on its side behind the front row), around the time Musk revealed the plans to build the new factory he Tweeted

Final decision made earlier this week on booster engine count. Will be 33 at ~230 (half million lbs) sea-level thrust. All engines on booster are same, apart from deleting gimbal & thrust vector actuators for outer 20.

That means every Starship and Superheavy combination will require 42 raptor engines: 33 in the Superheavy booster, plus three regular and three Vacuum Raptors in the Starship upper stage.  

While Raptor’s current design isn’t quite there, Musk says that SpaceX will debut an upgraded “Raptor 2” engine in the not too distant future, raising maximum thrust to 230 tons (~510,000 lbf). Aside from the removal of a few structural components required for engine gimballing on 20 booster Raptors, every engine on Starship – save for 3-6 vacuum variants – will thus be identical.

That's right; the 33 engines in Superheavy booster will produce 16.83 million pounds of thrust; over twice the thrust of the Saturn V that put Americans on the moon 50 years ago.

SpaceX photo of a sea-level Raptor (SN4) on the left and the Raptor Vacuum version in their Hawthorne, California facility. 

According to Musk, Raptor Vacuum production will remain in Hawthorne alongside work on “new, experimental designs.”


  1. Sounds like some serious upgrades for the ICBM inventory is about to happen.

  2. Absolutely amazing what the SpaceX is doing. And possibly up to 6 vacuum Raptors per Starship.

    Now, if only the FAA would get out of SpaceX's way.

  3. This makes me smile. It's the world I expected to live in many years ago. Silver Spaceships that take off and land vertically, and whole factories devoted to building nothing but ships and engines.

    Godspeed, Elon Musk!

    1. Yes, space as a routine thing, not as a hobby. Love it.

  4. Why gimbal any of the engines? If there is an outer ring of engines, why not just vary the thrust around the ring to get yaw and pitch? Add in some thrusters for roll control (the easiest axis to maneuver about) and you save a bunch of gimbals, flexible hoses, servos, etc.

    1. Gimballing is far easier than controlled differential thrust. And gimballing adds in the inherent vector change by pushing at an angle against the center spine of the space ship.

      Much like the efficiency of a center-line mounted rudder on a boat versus a steering oar. Push on the keel takes far less energy than pushing on the side.

      On small objects, like missiles, you can steer effectively with the steering vanes. But large Starship-sized objects, thrust vectoring is really the more efficient method. And actually reduces the complexity, as it's easier to gimbal an engine than it is to try moveable vanes or flaps on, yeah, a Starship-sized object.

      Thrusters are good for lower-speed motion control, or in-vacuum motion control, but aren't efficient at boosting through atmosphere. Thus, again, thrust-vectoring by gimballing.

      Everything is a trade-off. You could do vectoring using thrusters, but to effect the ship would require a lot of energy, which can be used more efficiently with a gimballed engine doing both direction change (at the end of the ship's rotation axis) and, well, boost-thrusting the whole thing.

  5. I knew he was going to settle on the magic number 42...

  6. This comment has been removed by a blog administrator.