Privately-funded rocket reaches orbit!

[KrazyHorse]

Basically, my guess is that there are two components to a satellite's construction costs:

1) Enough to get the job done
2) Insurance against launch costs in the form of additional reliability for components etc.

You can probably scale 2 pretty easily, but not 1. And concrete launch costs also don't take into consideration the cost of not having your sat working for 2 years before you can build and launch a replacement, so even if launches get dirt cheap you're going to build in additional reliability.

[Kuciwalker]

If launches get really cheap then it probably won't be a 2 year wait.

[KrazyHorse]

Not necessarily true. It takes a while to actually build the ****er, assuming you're not running a large constellation of identical satellites.

[DanS]

It stands to reason that satellite prices roughly scale with launcher prices. But space launch has limited suppliers and there are lots of gotchas (range availability and priority, manifest priority, ITAR, availability of rockets in a single weight class, etc.).

As an example, the Falcon 1 probably will be launching many of the Google Lunar X Prize competitors in the coming years. But you have to be first through to accomplish the task in order to get the prize. So some competitors probably will launch on other rockets, no matter the price of the probe and the price of the alternative launcher.

[KrazyHorse]

Again, I don't know if the scaling argument is true. The Iridium satellites have a construction cost of 5 mill per unit, apparently. I think it's fantasy to imagine that it could be done much cheaper. You're putting together a large (600 kg or so), complicated piece of machinery for operation in a hostile environment.

I don't think that taking launch costs from 40 mill to 8 mill will drastically alter that 5 mill figure.

[DanS]

I agree that it's not necessarily true. It would depend on the application, as demonstrated by the Google Lunar X Prize example.

But it stands to reason that as launch costs come down, engineers would optimize their machinery to take advantage of the new reality.

[Ben Kenobi]

10/1 x 1/4 = 10/4 = 2.25

Dammit. 5/2 = 2.5 Corrected. 9/4 is 2.25

[KrazyHorse]

I don't argue with that. But the question is how much scaling down they will be able to do.

[DanS]

My sense is that most applications will be able to scale a great deal and that new applications will become economically viable. That said, launch prices have never varied that much, so it's anybody's educated guess what the impact of a big change in price will be.

I think the change will be uneven and slow, as all of the factors that I mentioned above probably will change slowly. F.e., when you are building your satellite and doing a Falcon 1 v. Minotaur I trade, you have to consider the possibility that you will have to roll your payload from the Falcon 1 to the Minotaur I if SpaceX ceases operations or slots you on its manifest in a disadvantageous way. The satellite design optimizations for the Falcon 1 may be completely different than those for the Minotaur I.

[DanS]

Here anecdotal data for you, KH.

Avanti Communications of the UK is building the Hylas Ka-band communications satellite for roughly $170 million. They contracted with SpaceX for this launch on the Falcon 9 for probably about $40 million. Subsequent to signing on for the launch, it switched to the Ariane 5, which is priced at roughly $120 million.

[KrazyHorse]

Sorry, but what is that data supposed to tell me (not being snarky)?

[DanS]

The general prices of satellites in relation to launch costs. Not quite 1:1, but close.

It's my sense that the quoted Iridium costs, if correct, are atypical of the general satellite market.

[KrazyHorse]

Yes, because Iridiums were produced on an assembly line.

But you can't just list total cost to construct because that includes R&D. Second unit costs less.

And the replacement cost is what we're interested in. Specifically, because when a launch vehicle has a failure rate greater than 0 the expected loss from failure is the rate times the replacement cost.

[DanS]

I think you're making this analysis too complicated. Most people in commercial space aren't doing exotic R&D. They're contracting out to Hughes to build them a satellite with a certain number of transponders on a standard bus, etc. I doubt you'd get much of a discount on a replacement.

[Kuciwalker]

My understanding is that there are large fixed costs associated with making sure the thing doesn't fall apart during launch and making sure the thing doesn't fry in the solar radiation.

In particular, the "making sure it doesn't fall apart" has to be redone every time you change anything because they have to test for resonances and such.