Here's what "Rocket Boy" Homer Hickam had to say about the early days of rocket testing at what became NASA's Marshall Space Flight Center in Huntsville, Ala.: At Huntsville, ďthey had the solid example of the German rocket scientists who were extremely practical guys. They didnít believe in flying anything until they tested it until failure. They would run [rocket engines] until they blew up. And they knew exactly the parameters within [which] these engines worked, and what was fragile and what wasnít and what needed to be beefed up and what didnít. As versus today, they test it on a computer and they donít really know. You plug in a bunch of numbers, but you donít really know.Ē
Elon Musk is a remarkable man with a remarkable career behind him which continues to surprise us all. That said, I would take issue with the comment that "there are no prior precedents for success in creating an orbital space company, really". Actually from the earliest rocket launches, many countries have started up successful space launch programs. Most of these programs have a combination of government funding and private enterprise engineering effort. The evolutionary process continues.
This phrase is really key in my mind: "It is at the critical point of failure where the engineering lessons are found." I suspect that many people think of failure in terms of forgetting to tighten a bolt. In general, that sort of failure just tells you that you have poor workmanship and poor quality control.
The most useful type of failure comes from pushing a design well past its limits and seeing wear those limits are, or ensuring that the are far enough above the safety margin. When I hop on a Boeing 777, I don't want to know that the engineers designed it to withstand 150% of the maximum predicted load. I want to know that they stressed it until it failed and that the failure was passed the 150% design strength.
If the wing doesn't pass, then you get to see where and how it will likely fail and you can address that area. If it passes, then you've learned that it is as strong as you want it to be.
The latest on SpaceX launch to the International Space Station: NASA says the new launch date for the oft-delayed flight of the SpaceX Dragon cargo ship to the space station is now scheduled for May 19. Presumably, SpaceX engineers are still testing and verifying spacecraft software.
I of course was referring to the Dragon spacecraft, not the entire Falcon 9 stack.
We are also aware that there are many other commercial space companies that receive less attention than SpaceX. We have written about them here:
As an engineering website, we can widen out the question: and it comes down to the unglamorous arena of Testing.
Testing is where science meets engineering, and because of that you learn more from failure.
When your widget passes its stress test, you haven't learned what you need to know, which is how much lighter or cheaper you could make it. That saving could then be diverted into improving the weakest link in the system elsewhere.
Pushing test out to after launch, whether literally or not, is dumb and irresponsible engineering and bad science, no least because you can't instrument a disaster properly and you certainly can't repeat it at will.
When it fails you will have discovered what the margin of safety might be.
Testing is scientific research at its finest.
No-one said it isn't exciting flying by the seat of your pants, sure, and there are plenty of barnstormers who can't believe they are being paid to take risks
as test pilots. But no-one gives them passengers when they take up a new craft to plot its flight envelope.
'You never learn anything by being right'
Blog Doing Math in FPGAs Tom Burke 20 comments For a recent project, I explored doing "real" (that is, non-integer) math on a Spartan 3 FPGA. FPGAs, by their nature, do integer math. That is, there's no floating-point ...