[ansible]

It is a flawed comparison in the other way as well. The shuttles required vast amounts of refurbishment after each flight. While the thermal protection system was improved with the later shuttles, they required a lot of inspection and replacement of individual tiles, which was an enormous cost. It was also standard for the main engines (SSMEs) to be pulled out and swapped with ones that have been fully inspected. Even the solid rocket boosters (SRBs) needed to be towed back to port, disassembled, and completely refurbished for each flight.

The F9 first stage, on the other hand, typically goes through a relatively light amount of inspection and repair after each flight.

[TchoBeer]

Do you have specific numbers on how much inspection and refurbishment the space shuttle went through vs F9?

[ansible]

Per-flight costs for the shuttle was $1.6 billion USD (2010 dollars). A good chunk of that was refurbishment for the shuttle and the SRBs. I didn't find specific numbers in a quick Internet search.

https://www.space.com/12166-space-shuttle-program-cost-promi...

[no-s]

> Per-flight costs for the shuttle was $1.6 billion USD (2010 dollars).

That was the improved cost. In the late 80s I interviewed at the contractor running shuttle ops (Lockheed missiles & space?). The head of the NASA project office told me it was close to 2.4 giga$ to turn around a shuttle and they hoped to cure that with a new prime contractor arrangement (and also fix the management problem that was the proximate cause of the Challenger disaster, but apparently only suppressed it for a time...)

[reddog]

Isn't that the cost of the entire space shuttle program (including R&D) divided by the number of launches? I believe the real cost of turning around a shuttle flight towards the end to the program was a fraction of this.

I wonder what the cost of a currrent spacex flight is doing this same kind of accounting. I don't suppose anyone really knows since spacex is a private company.

[ansible]

In some sense, the Shuttle program was in continuous development. But to give you some idea of the work involved in even the later launches, consider this:

Instead of inspecting 24,000 tiles by hand, they developed a scanner to automate the process starting with STS-118:

https://www.nasa.gov/home/hqnews/2007/aug/HQ_07171_Shuttle_T...

They had hundreds of techs, working thousands of hours per launch to get each orbiter ready. SpaceX is expending a tiny, tiny fraction of that effort to get each stage-1 booster ready for re-flight. Part of that of course is that the booster is coming back at sub-orbital speeds.

So it is more fair to compare the F9 stage-1 to the pair of SRBs used for the Shuttle. But even then, there was a lot of effort just to get the SRBs ready for re-flight.