3 ms·

In any event, they’re going to need a better launchpad design. They’ll want a launchpad they can turn around and reuse rapidly in the future.

bryanlarsen 3y ago
According to Elon they started designing a water cooled plate to put under the launch pad 3 months ago.
- Zigurd 3y ago
  At this point there is project risk from them not wanting to redesign and rebuild the launch mount. If an add-on flame diverter is not enough, they will have just delayed a painful decision that imposes long lead-times. It might even be a blessing if the launch mount can't be repaired in a cost-effective way and they have to start over. But an iterative, Agile-like approach looks great when perfecting welded joints. But doing too many more iterations to get to orbit, in which boosters and Starhips are expended, could break the project. Agile is appropriate for a lot of software projects (n.b. not Scrum), but maybe not for the biggest rocket ever made.
  - bryanlarsen 3y ago
    If the flame diverter is not sufficient, it's likely that this launch will tell them that. So in that scenario, they've likely saved two months.
    - Zigurd 3y ago
      That is possible. But, so far, the decision-making has been faulty: the decision to go ahead without, at least, the flame diverter, was a poor decision. Elon will be loathe to make any decision other than "Patch it up and add a flame diverter that fits between the legs of the launch mount." Secondly, no decision at this point will save them two months. They are in for many months of reconstruction, apart from all the other design changes needed. They are also in for FAA scrutiny that will take at least several months. There is a large incentive to be unrealistic about this: A realistic view of delays makes an already very optimistic schedule for supplying a Starship-based lunar lander to the Artemis project look very risky. What we are seeing is the limits to scaling an iterative project. It is transitioning from learning how to weld stainless steel, which can evolve quickly, to learning how to fix multiple large problems when every iteration costs 39 rocket engines, 200 tankers of fuel/oxidizer, etc.

bryanlarsen 3y ago

According to Elon they started designing a water cooled plate to put under the launch pad 3 months ago.

Zigurd 3y ago
At this point there is project risk from them not wanting to redesign and rebuild the launch mount. If an add-on flame diverter is not enough, they will have just delayed a painful decision that imposes long lead-times. It might even be a blessing if the launch mount can't be repaired in a cost-effective way and they have to start over. But an iterative, Agile-like approach looks great when perfecting welded joints. But doing too many more iterations to get to orbit, in which boosters and Starhips are expended, could break the project. Agile is appropriate for a lot of software projects (n.b. not Scrum), but maybe not for the biggest rocket ever made.
- bryanlarsen 3y ago
  If the flame diverter is not sufficient, it's likely that this launch will tell them that. So in that scenario, they've likely saved two months.
  - Zigurd 3y ago
    That is possible. But, so far, the decision-making has been faulty: the decision to go ahead without, at least, the flame diverter, was a poor decision. Elon will be loathe to make any decision other than "Patch it up and add a flame diverter that fits between the legs of the launch mount." Secondly, no decision at this point will save them two months. They are in for many months of reconstruction, apart from all the other design changes needed. They are also in for FAA scrutiny that will take at least several months. There is a large incentive to be unrealistic about this: A realistic view of delays makes an already very optimistic schedule for supplying a Starship-based lunar lander to the Artemis project look very risky. What we are seeing is the limits to scaling an iterative project. It is transitioning from learning how to weld stainless steel, which can evolve quickly, to learning how to fix multiple large problems when every iteration costs 39 rocket engines, 200 tankers of fuel/oxidizer, etc.