4 ms·

It is baffling how many HNers in this thread don't understand what "test flight" means. Yeah, for SpaceX, experimental hardware is expendable. Yeah, it is not

It is baffling how many HNers in this thread don't understand what "test flight" means. Yeah, for SpaceX, experimental hardware is expendable. Yeah, it is not very typical in space industry. But it is a tried-and-tested approach, which gave them the Falcon 9, the most reliable and technologically advanced launcher of today.

preisschild 3y ago
Tbf we just saw SpaceX launch two rockets, which both failed, for 2 bln dollars per rocket
- inglor_cz 3y ago
  Yeah, but this is development of a brand new reusable superheavy launch system. Can't be cheap. We should keep that in mind and compare the comparable. Looking at the Western space industry history, Starship development is only comparable to the Apollo or the Space Shuttle program - both of which would be insanely expensive in 2023 dollars, certainly exceeding 100 billion USD each, perhaps 300 billion USD in case of Apollo.
- DylanSp 3y ago
  Got a citation on that $2 billion cost for each Starship launch? It's not clear what the exact cost is, but I haven't seen any estimates that are nearly that high.
gpm 3y ago
I don't have any problem with them adopting a test flight heavy methodology, however it is extremely misleading to suggest that this is the methodology which gave them the Falcon 9. The falcon 9 succeed on it's first launch, it was not a test flight heavy program. The falcon 1, SpaceX's only orbital rocket prior to the falcon 9, did fail on it's first three flights, however it was never intended to and that nearly bankrupt the company. They destroyed third party payloads on flights 1 and 3, because these were intended to work, not be test flights. Long after it was in production the falcon 9 did begin a "test-flight" heavy program, but that was for developing the ability to recover boosters after they had successfully delivered a customer payload in an operational flight, not for developing the rocket itself.
- inglor_cz 3y ago
  OK, I could have been more precise: the "hardware rich" program is what gave them the current reusable configuration of Falcon 9, Block Five. Which is what I would describe as the most reliable and technologically advanced launcher on the planet. The series of failures when attempting to land for the first time was spectacular.
  - JumpCrisscross 3y ago
    You also see dependencies early. Even if you never get to use the refuelling mechanism, you’ll gain data on how the reserved fuel sloshes and alters flight characteristics. This refines your models and may give clues as to efficiency gains that otherwise would have been an afterthought or, more likely, scrapped in the face of a deadline.
- sheepshear 3y ago
  All of the Falcon 1 payloads were expendable (which includes low-priority). It was a test vehicle conducting test flights in all but name. Had they continued developing it instead of pivoting to Falcon 9, it would have been its own test vehicle. Suppose that the eventual production vehicle has a fairly different design, a different name, and it succeeds on its first flight. Would it then have used the same methodology that resulted in Falcon 9? I'm only asking rhetorically. The Ship of Theseus is an interesting philosophical question, but the way you define "methodology" shouldn't depend on your answer.
godelski 3y ago
> It is baffling how many HNers in this thread don't understand what "test flight" means. I've been seeing parallels to this growing at an odd rate. Where context is removed and HN only cares about results at TRL ~7.5+. Its often quite baffling to see systems/models designed in academic spaces with minimal compute resources to systems/models with billions of parameters and millions of dollars in funding. Or to see TRL 2/3 projects be dismissed because they are not >TRL 6. I wonder how people that make these types of responses think research and development happens if you must catapult one's self from a theory to usable product in a realistic environment before it becomes of any interest. It seems rather anti-scientific tbh. Not what I'd really expect from a crowd dominated by developers, engineers, researchers, and scientists. As for rockets, well... exploding rockets are expected during testing phases. Obviously you want to reduce this, but catastrophic failures, or any failure, serve as great learning opportunities due to the inability to formulate a complete model (combustion instability is a notoriously difficult problem). I think also many are interested in the subject and so do watch videos and read other sources but that these are at high level discussions and often not getting into the low level nuance that is actually necessary for solving these types of problems. But that as humans we often conflate excitement and time with a subject with expertise in that subject (all time is not equal). Personally I wonder if this is an accurate and general phenomena and exacerbated by our growing information age combined with the large number of armchair science communicators. I think all nerds, including myself, have been guilty of asserting things as facts (or with stronger confidence than they should be), often in an effort to help inform others or simply wishing to participate. On a side note: anyone feel like it's a bit weird how low (aerospace) engineers are paid in comparison to us in software? Like not bad wages overall but there is a large gap (not looking to get my pay decreased, but the other way around) Blue Origin: 85k/119k/162k (bottom 10%/Median/Top 10%)(https://spacecrew.com/salaries/blue-origin-salary https://spacecrew.com/salaries/blue-origin-salary) SpaceX: 69k/112k/178k (https://spacecrew.com/salaries/spacex-salary https://spacecrew.com/salaries/spacex-salary)
- itsoktocry 3y ago
  >On a side note: anyone feel like it's a bit weird how low (aerospace) engineers are paid in comparison to us in software? Software produces gobs of money, the likes of which the world has never seen. Aerospace as an industry may not ever be profitable. It's as easy as that.

preisschild 3y ago

Tbf we just saw SpaceX launch two rockets, which both failed, for 2 bln dollars per rocket

inglor_cz 3y ago
Yeah, but this is development of a brand new reusable superheavy launch system. Can't be cheap. We should keep that in mind and compare the comparable. Looking at the Western space industry history, Starship development is only comparable to the Apollo or the Space Shuttle program - both of which would be insanely expensive in 2023 dollars, certainly exceeding 100 billion USD each, perhaps 300 billion USD in case of Apollo.
DylanSp 3y ago
Got a citation on that $2 billion cost for each Starship launch? It's not clear what the exact cost is, but I haven't seen any estimates that are nearly that high.

gpm 3y ago

I don't have any problem with them adopting a test flight heavy methodology, however it is extremely misleading to suggest that this is the methodology which gave them the Falcon 9. The falcon 9 succeed on it's first launch, it was not a test flight heavy program. The falcon 1, SpaceX's only orbital rocket prior to the falcon 9, did fail on it's first three flights, however it was never intended to and that nearly bankrupt the company. They destroyed third party payloads on flights 1 and 3, because these were intended to work, not be test flights. Long after it was in production the falcon 9 did begin a "test-flight" heavy program, but that was for developing the ability to recover boosters after they had successfully delivered a customer payload in an operational flight, not for developing the rocket itself.

inglor_cz 3y ago
OK, I could have been more precise: the "hardware rich" program is what gave them the current reusable configuration of Falcon 9, Block Five. Which is what I would describe as the most reliable and technologically advanced launcher on the planet. The series of failures when attempting to land for the first time was spectacular.
- JumpCrisscross 3y ago
  You also see dependencies early. Even if you never get to use the refuelling mechanism, you’ll gain data on how the reserved fuel sloshes and alters flight characteristics. This refines your models and may give clues as to efficiency gains that otherwise would have been an afterthought or, more likely, scrapped in the face of a deadline.
sheepshear 3y ago
All of the Falcon 1 payloads were expendable (which includes low-priority). It was a test vehicle conducting test flights in all but name. Had they continued developing it instead of pivoting to Falcon 9, it would have been its own test vehicle. Suppose that the eventual production vehicle has a fairly different design, a different name, and it succeeds on its first flight. Would it then have used the same methodology that resulted in Falcon 9? I'm only asking rhetorically. The Ship of Theseus is an interesting philosophical question, but the way you define "methodology" shouldn't depend on your answer.

godelski 3y ago

> It is baffling how many HNers in this thread don't understand what "test flight" means. I've been seeing parallels to this growing at an odd rate. Where context is removed and HN only cares about results at TRL ~7.5+. Its often quite baffling to see systems/models designed in academic spaces with minimal compute resources to systems/models with billions of parameters and millions of dollars in funding. Or to see TRL 2/3 projects be dismissed because they are not >TRL 6. I wonder how people that make these types of responses think research and development happens if you must catapult one's self from a theory to usable product in a realistic environment before it becomes of any interest. It seems rather anti-scientific tbh. Not what I'd really expect from a crowd dominated by developers, engineers, researchers, and scientists. As for rockets, well... exploding rockets are expected during testing phases. Obviously you want to reduce this, but catastrophic failures, or any failure, serve as great learning opportunities due to the inability to formulate a complete model (combustion instability is a notoriously difficult problem). I think also many are interested in the subject and so do watch videos and read other sources but that these are at high level discussions and often not getting into the low level nuance that is actually necessary for solving these types of problems. But that as humans we often conflate excitement and time with a subject with expertise in that subject (all time is not equal). Personally I wonder if this is an accurate and general phenomena and exacerbated by our growing information age combined with the large number of armchair science communicators. I think all nerds, including myself, have been guilty of asserting things as facts (or with stronger confidence than they should be), often in an effort to help inform others or simply wishing to participate. On a side note: anyone feel like it's a bit weird how low (aerospace) engineers are paid in comparison to us in software? Like not bad wages overall but there is a large gap (not looking to get my pay decreased, but the other way around) Blue Origin: 85k/119k/162k (bottom 10%/Median/Top 10%)(https://spacecrew.com/salaries/blue-origin-salary https://spacecrew.com/salaries/blue-origin-salary) SpaceX: 69k/112k/178k (https://spacecrew.com/salaries/spacex-salary https://spacecrew.com/salaries/spacex-salary)

itsoktocry 3y ago
>On a side note: anyone feel like it's a bit weird how low (aerospace) engineers are paid in comparison to us in software? Software produces gobs of money, the likes of which the world has never seen. Aerospace as an industry may not ever be profitable. It's as easy as that.