No Substitute for Trial-and-Error

Congratulations to SpaceX on reaching orbit with the Falcon 1 some months ago!

Since then, my elation about the success has passed and a more earthly assessment of the implications has started to take shape.  As long as it continues to launch satellites successfully, SpaceX will start driving down the price of launch overall.  However, the fall in prices may be slower than we hope because SpaceX probably will have more failures.  Perhaps soon.  Perhaps on the next launch of the Falcon 1.  In a more perfect world, we would hope that the failures happen sooner rather than later so that any problems in the design, fabrication, and launch of these rockets can be hammered out.

It has come as something of a revelation to me that trial-and-error cures most ills.  Even bad designs sometimes can be made to fly, given enough trial-and-error.  The trick is to have enough bankroll to survive the process.  You take a look at your bankroll and pick your process such that you don't run out of money doing the repetitions.  The US Government has almost an unlimited bankroll and therefore uses a spectacularly expensive process.  Elon Musk had several hundred million dollars in the bank and picked a focused group of a couple hundred employees.  John Carmack had maybe ten million dollars in the bank and picked a focused group of ten volunteers.  Korolev had his bankroll and process too.

Likewise, even great rockets will have failures in testing.  I think that the Falcon 1 & 9 are beautiful and relatively simple designs.  The company is using the latest in manufacturing technology.  And launch operations are lean and professional.  But even so, the company had three failures before a success.  The Falcon 9 probably will have at least one failure early on.  Hopefully, Musk has a bankroll equal to the task.  It seems likely that he is almost out of ready cash, what with Tesla Motors needing more equity investment than originally planned and Musk's recent divorce announcement.

Lately, I have been following Armadillo Aerospace and feel the most comfortable with its approach to design and testing.  The number of flight reps that Armadillo performs on its hardware is impressive.   Failures don't lead to existential questions for the company, unlike with SpaceX.  Carmack's bankroll appears more than up to the relatively modest task of sustaining development and failure.  The approach is not without its pitfalls, of course.  Some followers or supporters of the company may become impatient.  Carmack may lose interest or -- God forbid -- he may otherwise become unable or unwilling to push the company forward.

At the end of the day, the reward in systems that are sustainable favors Armadillo hugely.  An Armadillo manned orbital system definitely would be cheap enough to serve average people like me.  This is what I keep my eye on.  Further, I wonder if the risk of project failure is equal or lower for Armadillo versus the NASA programs or even SpaceX's programs.  After all, if Carmack runs out of money, at least there are many more people available with the capability to fund the project in the breach.  If Musk runs out of money, only billionaires will be able to help him.  If NASA runs out of money, the project is shut down.

Comments

[David C. Neal]

I have never understood why SpaceX did not
go the route of flying the Falcon 1 with a dummy upper stage, just to gain experience
with flight ops and to have a chance to
perfect the stage 1 recovery. Of course, there
are monetary concerns to this approach, but
customer confidence has a high potential
value as well. I wish them success, and I hope eventually they can operate more by
demonstrated ability than luck.

[Jason Gerend]

I agree that a more incremental testing approach yields a more thoroughly tested vehicle. However, I think that in SpaceX’s case there may not have been a lot of added value in flying the first stage repeatedly on its own. If the second stage development delayed testing of the first stage, then perhaps, but I’m not sure that it did. If the goal is to demonstrate how reliable your vehicle is, then flying a bunch of dummy full-up missions would be a solution, but I think that SpaceX is acquiring enough demand without spending excess capital on test missions. From a business perspective, they should do as few unpaid flight tests as possible while still getting enough reliability to retain their customers. It might not yield as fast a development plan as if you tested a lot, but it’ll use up your capital slower…

[Robert Horning]

From what I understand about SpaceX, the reason they had actual payloads for flights 1-3 of the Falcon 1 was due to some huge discounts to the customers, on the presumption that it was going to be an early test flight that did at least have the potential of making it into space.

The Air Force Academy satellite was one of those project where paying for a full-priced proven launch was just a little bit high, but paying half price (I don't know the exact SpaceX figure, but it was a reduced rate) for something that had a 50% chance of success was considered a reasonable gamble.

SpaceX didn't bother with a paying customer on flight #4 due to the short period of time between flights #3 and #4. SpaceX also needed desperately to prove that they could actually get something "up there", so Elon Musk wasn't going to take any other chances with an extra payload. Even in this case, SpaceX had to manufacture a "simulated" payload at the last minute... and what better "simulated" payload could there be than a real life project that would have some value in space.

In this, I don't fault SpaceX for at least trying, and it wasn't like there was anybody harmed with a failure... where the customers involved in the earlier tests knew full well that a failure was not only possible but likely.

[Daniel Schmelzer]

I think that SpaceX didn't do launch testing of the lower stage for strictly business reasons. Musk saw a window of opportunity from a marketing standpoint and was willing to sacrifice some failures in the future to hit the window now.

Looking at his launch manifest, it's tough to gainsay his judgment. But it doesn't mean that those failures in the future will go away.

[Juicy]

I would name your blog the dreamland! While Santa knocks at our door just once per year, you blog is open the whole year – wow!
http://www.juicy-couture-cheap.com/

[POLO]

I would name your blog the dreamland! While Santa knocks at our door just once per year, you blog is open the whole year – wow!
http://www.cheaprlpoloshirts.net/

[poloralphlauren]

Thank you for this article. That's all I can say. You most definitely have made this blog into something special. You clearly know what you are doing, you've covered so many bases.Thanks!

http://www.discountpoloralphlauren.org/

[Arena animation Bangalore]

I love learning from others and this article shows me how much I still have to learn. Thanks for this wonderful post.

[hot sell nike af1]

The decline of literature indicates the decline of a nation ; the two keep in their downward tendency.