Transterrestrial Musings




Defend Free Speech!


Amazon Honor System Click Here to Pay




Site designed by


Powered by
Movable Type 4.0
Biting Commentary about Infinity, and Beyond!

« Flash Boot? | Main | Off to California »

Initial Thoughts

It's looking like there was a second-stage problem, either separation, or ignition (or both, since one could cause the other).

As I said last night, this is obviously a disappointment to the SpaceX team. Particularly since they had previously had a flight where this wasn't a problem, so in a sense it was two steps forward, one step back. I think that at this point, almost anyone is going to be pretty leery of putting a payload on the vehicle until it's had at least one successful flight. Is it the end? Despite what Elon said a long time ago about three strikes, it's hard to see it now. He's fully invested now, both financially and (I would imagine) emotionally, and he's not going to come this far just to give up, particularly when tantalized by his previous almost-success on the second flight.

They'll go through the telemetry, figure out as best they can what happened, and try again, and hopefully soon. In a sense, as someone noted in comments in the earlier post, Falcon 1 is really a test program for the bigger vehicles, though they should get an operational small launcher out of it as well.

As always, this points up the problem with expendable vehicles. They are very expensive to flight test, so you can't afford to do very many, and every flight is a first flight, so you can't wring bugs out of a vehicle with incremental testing. And it's a lot harder to figure out what went wrong because you generally don't get much debris to analyze (the first flight that failed off the pad was a rare exception)--you have to dig through electronic entrails. And NASA, of course, in its cargo-cult determination to redo Apollo, is taking exactly the same expensive and unreliable approach.

And just checking now, I see that Clark is having similar thoughts to mine.

Once the problem that caused this failure is determined, I would suggest that SpaceX just bite the bullet and allocate 2 or 3 Falcon I vehicles for test flights and fly them within a relatively short period, say six months.


This would represent a $20M-$30M investment but until the Falcon I is flying reliably, SpaceX will find it very difficult to get any more commercial or government payload contracts and it won't have any chance of getting COTS D (ISS crew transport) funding. The Falcon 9 is a completely different vehicle but the Falcon I is what currently defines the company's ability, or inability, to deliver what it says it can.

Anyway, best of luck to them in the future, but they know that they need more than just luck.

[Update a couple minutes later]

I see that Elon has a statement, which confirms my suspicion above:

There should be absolutely zero question that SpaceX will prevail in reaching orbit and demonstrating reliable space transport. For my part, I will never give up and I mean never.

That's the kind of attitude you have to have, even if eventually, you do in fact have to give up. I hope he won't have to.

Also note Clark's comment at the end of the post, that SpaceX is following in the tradition of all expendable staged launch vehicles in its failure modes, though they do seem to be getting the avionics right.

 
 

0 TrackBacks

Listed below are links to blogs that reference this entry: Initial Thoughts.

TrackBack URL for this entry: http://www.transterrestrial.com/admin/mt-tb.cgi/10032

20 Comments

Tom Hill wrote:

My first thought is they'll need to look at what they re-did after last flight's re-contact between stage 1 and 2. Look at what you changed first...

Very happy to see that Elon said he's not giving up. I am curious what the investment he mentioned is, though...

Bill White wrote:

Why would a non-expendable vehicle be less prone to catastrophic failure during its initial test flights?

This stage separation failure also would have destroyed a two-stage-to-orbit (TSTO) RLV payload, right?

Rand Simberg wrote:

Why would a non-expendable vehicle be less prone to catastrophic failure during its initial test flights?

Because it wouldn't try to go all the way to orbit on its first flight. It would have much smaller, incremental tests, gradually expanding the envelope. The stage separation mechanism would probably be redundant, and tested with smaller upper stage simulators that could be returned to the launch site in an abort of the separation failed.

Bill White wrote:

The stage separation mechanism would probably be redundant, and tested with smaller upper stage simulators that could be returned to the launch site in an abort of the separation failed.

Can you explain how a stage separation mechanism can be redundant? It also seems to me the stages either separate or they do not separate, whether its an expendable upper stage or a reusable upper stage.

What am I not seeing?

That said, repeated flight testing before adding genuine customer payloads is a good idea (if the funding is there). And that looks to be what Musk shall do.

Habitat Hermit wrote:

Bill they either separate or don't (well... they can partially separate too or separate while introducing or contributing to errors which is what happened in flight 2) but they can do it under different conditions/circumstances and thus with different results. Those conditions (and thus results) can be incrementally expanded from where the consequences are unlikely to result in the loss of vehicles/launches.

I'm also curious about the investment, not many people talking about it yet but in some ways it's bigger news than the failure.

Anyone know of a good historical account/summary of early (relative to each individual launch system) launch failures? Maybe it hasn't been written yet --any takers? ^_^

Mike Puckett wrote:

If I were to guess who the investor is, I would guess its Bigelow.

He has a lot riding on the success of SpaceX.

Anon. wrote:

Bigelow has been adamantly opposed to investing in any launch operation. Can't see any evidence that has changed recently. Another guess would be Northrop Grumman (they have no currently dog in the hunt for ELVs) or maybe the guys at Google.

Paul Breed wrote:

From the book.... atlas the ultimate weapon

First flight 24.7 seconds to tumble.
2nd flight 32.6 seconds to tumble.
3rd flight a success.

Of the first 8 series A
2 fail,
3 partial succes
and 3 fully successful.

Of the next 10 series B
1 Fail
4 partial success
5 full success

Of the 7 Series C
1 exploded.
two are partial successes
and four were complete successes.

Series D (Operational)
36 flights with 29 successes.

Series E and F
Of 18 E flights
10 were successfule
Of 13 F flights 9 were successful.

So out of the first 92 flights 60 were complete successful. 65 %



ken anthony wrote:

Elon has said he plans to take Spacex public at some point. It's not unusual for a company that intends to go public to first get private investors before the IPO.

It IS big news and suggests that a public offering is likely once they have a record that would support an IPO.

When asked, "How much are you worth?" Elon's reply was, "that depends on the value of Spacex." or something like that. If he can undercut the competition by 10 to 50 percent that's worth billions I would think.

I don't see Bigelow as the investor. It doesn't even have to be anyone in the aerospace field. It's the potential numbers that matter to an investor.

every flight is a first flight, so you can't wring bugs out of a vehicle with incremental testing.

The first part is literally true, but I think the second part is less true. Each flight is an incremental test in a sense. If manufacturing is sound, identical vehicles should have the same performance which is not true of a reusable which will have wear (with refurbishing) from one flight to the next. What's incremental in a reusable is the cost, since you're not throwing everything away each flight. Unless I'm misunderstanding your point (now that could never happen!)

OTOH, wring out is probably the operative phrase because it is much easier to evaluate something when you can bring it back and look at it.

Ed Minchau wrote:

"...SpaceX will find it very difficult to get any more commercial or government payload contracts..."

That isn't necessarily so. There is one area of the launch services market that would provide a lot of payloads: cubesats. These types of satellites are a standard size (10cm cube) and mass (1kg) and fit into a standardized launch cradle. Such a satellite can be built by high school students, and high schools could probably afford to launch them at the going per-kg rate - it would likely be about 15 grand each or less as the flight rate increases.

That would open up space to high schools and to any small business that wanted a tiny satellite in orbit. Look at what happened when computers became accessible in price to schools and businesses; it's what gave us widespread use of the internet, and much less expensive computers.

The keys are reliability and high flight volume. If SpaceX starts launching a rocket a day, with a few hundred cubesats on each flight, then they can incrementally improve their flight operations and rocket designs and construction processes. The only way to get real reliability is the incremental testing. Look at what Scaled did with SS1: they slowly increased the envelope.

The big thing is to get a lot of flights up and incrementally improve. Quantity has a quality all its own.

Bill Hensley wrote:

The big question is how their customers will react. It does seem unlikely that anyone would be willing to put their payload on a Falcon 1 until they'd seen at least one successful flight. Preferably two. I agree that Elon is probably going to need to eat the cost of a couple of test launches to win back the confidence of his market. It's good to hear they still have a strong cash position thanks to the new investment. Long term I still wouldn't bet against them.

kert wrote:

Im just catching up with the news, but was the vehicle destructed ? No first stage recovery again ?

John Gebbert wrote:

"Long term I still wouldn't bet against them."

I would. The odds are against them. They've made it clear that they cannot achieve substantial cost reduction without reusability, and nobody else has done reusability yet, so there is no proof that it can be made to work. That's a whole other step beyond simply getting a successful launch vehicle.

But the big issue is that in the launch business, the most important thing to customers is reliability. Launch is usually a small percentage of their overall cost, so a reduction in launch cost is not important unless it is huge. Reliability is far more important. A traditional customer is going to avoid SpaceX because of the reliability issue. In fact, even a free launch is not attractive if it is high risk. Can a company like XM Radio afford to lose hundreds of millions of dollars on a satellite and revenue even if the launch is free? (No.)

So now Musk has to demonstrate reliability, and probably on more than one launch. Then he has to find some customers. And in order to be competitive long term, he has to do something that nobody else has done yet, which is recover and reuse his first stage. Lots of "ifs" in that equation, and therefore I would bet against them even if Musk chooses to stay in the business for the next several years.

ken anthony wrote:

he has to...

...recover and reuse his first stage

No, he doesn't. But if he does, it's gravy and gives him a substantial cost advantage over his competitors.

...find some customers

Since he's already profitable I think he's already proven that ability.

...demonstrate reliability

True. I'm ready to place my bets. They've only had 3 launches. They almost had it on there 2nd attempt. That is damned impressive.

...probably on more than one launch

There's a joke in there somewhere???

I would bet against them even if Musk chooses to stay in the business for the next several years.

I'll take that bet. How much?

John Gebbert wrote:

"No, he doesn't. But if he does, it's gravy and gives him a substantial cost advantage over his competitors."

He does. How is he going to compete against Boeing, or Arianespace, or ILS, if all he offers is a slight cost advantage and a largely unproven rocket?

"...find some customers

Since he's already profitable I think he's already proven that ability."

He's profitable? Can you explain how he has not delivered a single payload to orbit and yet is making a profit?


"True. I'm ready to place my bets. They've only had 3 launches. They almost had it on there 2nd attempt. That is damned impressive."

Are you aware of launch statistics? Three failures in a row is not impressive.

"...probably on more than one launch

There's a joke in there somewhere???"

Do you know how the business works? Most commercial customers, like comsat companies, look at reliability for the past several launches, usually 7-10. A single launch success, after three failures, does not get SpaceX out of the woods.


"I would bet against them even if Musk chooses to stay in the business for the next several years.

I'll take that bet. How much?"

$100 says that the company is not profitable by 2010.

Bill Hensley wrote:

I said I wouldn't bet against them because, if they have deep enough pockets, they will get the bugs worked out and then they will still have a simpler rocket with lower labor costs than the competition, which means a significantly lower cost structure. Musk has consistently said he is not counting the potential savings of reusability in his pricing. If they can achieve reusability(of which I am also skeptical) then, yes, it is gravy.

Having said all that, there's no doubt that this is a significant setback for SpaceX and their biggest problem will be regaining their customer's confidence. I say regaining, because in case you hadn't noticed they have (had?) a significant number of launches on their manifest. They are probably going to have to eat the cost of a couple of test launches, at least, before folks like the Malaysians are going to want to send their bird up.

Prior to this point Musk has said they were profitable already. That's a different thing from saying they are cash flow positive. U.S. accounting rules for revenue recognition using the accrual system say that you accrue revenue when the contract is booked, even when you haven't collected the money yet. Of course, if customers start canceling contracts then the situation changes.

ken anthony wrote:

$100 says that the company is not profitable by 2010.

Well John, you've just lost a bet as Bill just explained to you. I'll tell you what, send the $100 to Rand as a donation and we'll call it even.

If Spacex is able to achieve orbit with it's next 3 flights they still are not out of the woods. If they achieve orbit, then lose another, it doesn't mean they're over.

Do you know how the business works?

Yes. I do. Now honor your bet and pay Rand the money.

ken anthony wrote:

I have a question that's been on my mind since hearing that the problem was a separation issue. I thought the flight only lasted for about 1 minute 40 seconds, but the 1st stage is supposed to fire for about 3 minutes. Anyone have any ideas about that, or am I not seeing something?

Bill Hensley wrote:

Ken, the webcast video stopped about 2:11 into the flight, but I think they had at least a 30 second delay on the feed. That would put the "anomaly" at about 2:40, which is about when staging was supposed to happen.

Habitat Hermit wrote:

Thank you Paul Breed that was exactly the kind of thing I was looking for. I think I'm going to try to find a good clean source of such data for other launch systems as well or maybe piece one together.

Ken Anthony wrote:
"The first part is literally true, but I think the second part is less true. Each flight is an incremental test in a sense."

I'm prone to thinking that as well, if we're missing something it might be the scope and resolution of "incrementalism": it probably is a much smoother and simpler curve for reusable vehicles and a surprisingly jagged graph for expendable ones. Wear and tear from reuse is part of the incremental approach (feedback).

Ed Minchau the payload capacity of Falcon 1 is about 670 kg to LEO. Let's give 100 kg to various on-orbit payload support and launch/release systems and we get about 57 cubesats in one go and in pretty much the same orbit/inclination/altitude. Not impossible but I would think there are lots of different issues with that which bring prices up and/or could bring perceived value and applicability down.

Leave a comment

Note: The comment system is functional, but timing out when returning a response page. If you have submitted a comment, DON'T RESUBMIT IT IF/WHEN IT HANGS UP AND GIVES YOU A "500" PAGE. Simply click your browser "Back" button to the post page, and then refresh to see your comment.
 

About this Entry

This page contains a single entry by Rand Simberg published on August 3, 2008 5:58 AM.

Flash Boot? was the previous entry in this blog.

Off to California is the next entry in this blog.

Find recent content on the main index or look in the archives to find all content.

Powered by Movable Type 4.1