Author Topic: Why is space so expensive  (Read 18708 times)

Online john smith 19

  • Senior Member
  • *****
  • Posts: 6159
  • Everyplaceelse
  • Liked: 814
  • Likes Given: 5105
Re: Why is space so expensive
« Reply #40 on: 09/10/2017 12:55 AM »
Yes and no.

If you build and test them for the expectation of 5-15 years of on orbit use, they are expensive.

If you build/test for 1-3 years of on orbit use (remember that current rate of technological obsolescence is about 1.5 years), they are relatively inexpensive.

Note that there is a cost inversion underway. The sat constellation isn't that much more expensive than the ground based one, if we keep refreshing it for obsolescence.
With that kind of lifespan you can run with no redundancy single string systems, eliminating redundancy and cross strapping logic and control, either saving weight or adding capacity.

Provided your launch costs are low enough and your capacity growth high enough to support the continual  upgrading and redesign effort needed. IOW rolling innovation, where the design of the next generation starts almost before the last has completed deployment.

But note, this sort of break neck innovation can only continue while you have complete control of the system, end to end.  :(
Once it has to interface to outside standards that part of the system will be constrained to that standard. That part of the system can only innovate at the speed that that standard innovates at.  :(
Time will tell how serious an issue this is with "MuskNet."
"Solids are a branch of fireworks, not rocketry. :-) :-) ", Henry Spencer 1/28/11  Averse to bold? You must be in marketing."It's all in the sequencing" K. Mattingly.  STS-Keeping most of the stakeholders happy most of the time.

Online Space Ghost 1962

  • Senior Member
  • *****
  • Posts: 2474
  • Whatcha gonna do when the Ghost zaps you?
  • Liked: 2375
  • Likes Given: 1736
Re: Why is space so expensive
« Reply #41 on: 09/10/2017 06:09 PM »
Yes. As far as the business logistics equivalence, one location (in this case Vandenberg) preempts all global locations (after you've launched all planes of your constellation).

That is the power of a space based system.

I don't understand this.

When you plan a ground based communications/navigation system, there are specific locations to deploy resources/"assets" to, that allow you to service areas or have the logistics to allow service with. You buy/barter to gain them, and you refresh the technology based on the costing to phase in/out waves as you can afford to do so.

With telcos, they'd think in terms of decades (at one time when it was the Bell System they were thinking of 50 year asset/location cycles!). One of the reasons that it cost $384B to do third gen cell technology was to write off the second gen, buy/deploy the third, and enter service across the service area (with AT&T, they consolidate two different networks - "orange" and "blue").

For that amount of capital, you could constantly refresh a constellation, getting your core/edge upgraded the same and just concurrently cycle (serpentine) through an architecture as it goes 10x, 100x, 1000x in aggregate bandwidth (instead of the piddling 20x of the 2nd-3rd gen step up).

And the only location that has to matter is the one that everything goes through - the constellation launch location.

Earthbound businesses always think of "locking up" locations. (A recent presentation by IBM at Facebook's "Video at Scale" conference was a resurrected pitch from the 2000's about doing just this, with the intent to do an entirely ground based, massive deployment of optical switching with CDNs to "lock out" LEO constellations.)

The only way to compete with constellations are with a overwhelming technology on orbit approach that doesn't care about long term reliability (because you keep on shoveling sats when needed), but does care about maximum yield (aggregate bandwidth) from the constellation, where you assume sat/launch costs are cheap/disposable. If you get them wrong it doesn't matter because the next salvo does better and eventually you win.

Just so long as you keep launching from that location. 

Offline gongora

  • Global Moderator
  • Senior Member
  • *****
  • Posts: 2092
  • US
  • Liked: 1457
  • Likes Given: 1036
Re: Why is space so expensive
« Reply #42 on: 09/10/2017 06:42 PM »
Just so long as you keep launching from that location.

It makes more sense when you write more than one sentence  ;)  Although I'll add my normal reminder that the SpaceX constellation isn't polar and a majority of it doesn't need to launch from Vandenberg.

edit: it also remains to be seen how reliable satellite internet will be at different frequencies as they move more to Ka and V bands.
« Last Edit: 09/10/2017 06:43 PM by gongora »

Online Space Ghost 1962

  • Senior Member
  • *****
  • Posts: 2474
  • Whatcha gonna do when the Ghost zaps you?
  • Liked: 2375
  • Likes Given: 1736
Re: Why is space so expensive
« Reply #43 on: 09/10/2017 07:00 PM »
Just so long as you keep launching from that location.

It makes more sense when you write more than one sentence  ;)
What? And lose my "crisp-ness"? (Execs, investors, and command require such.)

Quote
  Although I'll add my normal reminder that the SpaceX constellation isn't polar and a majority of it doesn't need to launch from Vandenberg.

edit: it also remains to be seen how reliable satellite internet will be at different frequencies as they move more to Ka and V bands.
Back to "blather" mode.

Now we dig deeper into different approaches to using constellations (sometimes changes/changed with plans too).

Note that when you have short enough links, you can use retransmission to save you if the drops have small enough frequency. The point is to overwhelm with technology, fix up as needed, flood with volume, and to not look back. Or wait to have the perfect, stable, thing - because technology moves too fast.

Offline MikeAtkinson

  • Full Member
  • ****
  • Posts: 1684
  • Bracknell, England
  • Liked: 454
  • Likes Given: 53
Re: Why is space so expensive
« Reply #44 on: 09/10/2017 09:09 PM »
With telcos, they'd think in terms of decades (at one time when it was the Bell System they were thinking of 50 year asset/location cycles!). One of the reasons that it cost $384B to do third gen cell technology was to write off the second gen, buy/deploy the third, and enter service across the service area (with AT&T, they consolidate two different networks - "orange" and "blue").

For that amount of capital, you could constantly refresh a constellation, getting your core/edge upgraded the same and just concurrently cycle (serpentine) through an architecture as it goes 10x, 100x, 1000x in aggregate bandwidth (instead of the piddling 20x of the 2nd-3rd gen step up).

Deploy for USA (and/or Europe) and get the rest of the world for "free".

There is also the potential for a single solution for both fixed and mobile, something that Fibre or cellular systems do not give. Probably a few generations away, the technology to cover high density areas and reach into buildings from space is not available yet.

Offline tdperk

  • Full Member
  • **
  • Posts: 243
  • Liked: 73
  • Likes Given: 24
Re: Why is space so expensive
« Reply #45 on: 09/10/2017 10:33 PM »
Dead wrong and total nonsense.  Ed nailed it.    Don't need these types of posts.  It is much like political parties talking about each other.  A lot of rhetoric and no real truth.

Excerpt from Elon Musk biography:

Quote
"In addition to building its own engines, rocket bodies, and capsules, SpaceX designs its own motherboards and circuits, sensors to detect vibrations, flight computers, and solar panels," Vance wrote. "The cost savings for a homemade radio are dramatic, dropping from between $50,000 to $100,000 for the industrial-grade equipment used by aerospace companies to $5,000 for SpaceX's unit."

So I stand by my statement. The question is "Why is space so expensive" and I do not think the answer would be complete without mentioning this aspect. You can call it rhetoric, but it needs to be mentioned.

It's a lot cheaper to pay your own staff to do R&D and low rate production than buy it from someone else. Assuming your company has the resources to do the work in house, which can cost a lot of money. I bet the $5,000 per unit cost doesn't include the investment to be able to do the work at SpaceX.

And I bet the cost of the investment to do it in house was less than the cost of paying someone for else's ROI on that investment with their golden parachute given priority.

Vertical integration is a chief means by which SpaceX has made certain of lower cost--not lower risk, cost.  They had lower risk because for the first time the government was prepared to obey the laws against it's pet contractors competing against commercial ventures.

Once the government was out of the way, that market forces would lower the cost of access to LEO was a certainty.

There has never been a physical reason for orbiting a pound to cost more than a low two digit multiple of the energy cost to do so for several hundred pounds worth, or for bulk access to be more than tenfold the energy cost to do so.

That the cost of space access stayed north of $5k/lb for the 60 some years after the space age was inaugurated is a matter of the result of political policy, not physics or our relative inability to calculate how to cut available materials to accomplish far lower cost space access.
« Last Edit: 09/10/2017 10:44 PM by tdperk »

Offline tdperk

  • Full Member
  • **
  • Posts: 243
  • Liked: 73
  • Likes Given: 24
Re: Why is space so expensive
« Reply #46 on: 09/10/2017 10:40 PM »
Dead wrong and total nonsense.  Ed nailed it.    Don't need these types of posts.  It is much like political parties talking about each other.  A lot of rhetoric and no real truth.

Excerpt from Elon Musk biography:

Quote
"In addition to building its own engines, rocket bodies, and capsules, SpaceX designs its own motherboards and circuits, sensors to detect vibrations, flight computers, and solar panels," Vance wrote. "The cost savings for a homemade radio are dramatic, dropping from between $50,000 to $100,000 for the industrial-grade equipment used by aerospace companies to $5,000 for SpaceX's unit."

So I stand by my statement. The question is "Why is space so expensive" and I do not think the answer would be complete without mentioning this aspect. You can call it rhetoric, but it needs to be mentioned.
You do know that SpaceX has acknowledge that it spent, or is spending, a cool $1 billion just to develop first stage recovery?   - Ed Kyle

No Ed, that R&D cost is for the all their operations up to the point they were recovering the 1st stage.

And at that point they would have already cleared a net profit with operating costs.
« Last Edit: 09/10/2017 10:40 PM by tdperk »

Offline RedLineTrain

  • Full Member
  • ****
  • Posts: 537
  • Liked: 260
  • Likes Given: 247
Re: Why is space so expensive
« Reply #47 on: 09/10/2017 10:42 PM »
You do know that SpaceX has acknowledge that it spent, or is spending, a cool $1 billion just to develop first stage recovery?  Given that number, it seems likely that Falcon 9 itself has and is costing multiple billions of dollars to develop.  Don't kid yourself about what this stuff costs.  SpaceX may have found a more cost-efficient path to develop its big rockets than others, but it still isn't cheap!

Judging by the Air Force AR1 and Raptor contracts, traditional staffing an engine R&D engine development costs about a half order of magnitude more than lean staffing.  Of course, there's a difference in what the Air Force will obtain:  well understood technology main engine v. novel technology sub-scale upper stage engine with potential other uses.

It seems possible to see such R&D cost differences between vendors on other, similar projects.
« Last Edit: 09/10/2017 10:50 PM by RedLineTrain »

Offline gongora

  • Global Moderator
  • Senior Member
  • *****
  • Posts: 2092
  • US
  • Liked: 1457
  • Likes Given: 1036
Re: Why is space so expensive
« Reply #48 on: 09/10/2017 10:50 PM »
Judging by the Air Force AR1 and Raptor contracts, traditional staffing an engine R&D engine development costs about a half order of magnitude more than lean staffing.  Of course, there's a difference in what the Air Force will obtain:  well understood technology main engine v. novel technology sub-scale upper stage engine.

SpaceX was building Raptor anyway and asked for a much smaller percentage of the development cost from the government.  (It also has a potentially useful side effect of giving the government some insight into the development if SpaceX wants to get a Raptor based vehicle certified later).

Offline RedLineTrain

  • Full Member
  • ****
  • Posts: 537
  • Liked: 260
  • Likes Given: 247
Re: Why is space so expensive
« Reply #49 on: 09/10/2017 10:53 PM »
Judging by the Air Force AR1 and Raptor contracts, traditional staffing an engine R&D engine development costs about a half order of magnitude more than lean staffing.  Of course, there's a difference in what the Air Force will obtain:  well understood technology main engine v. novel technology sub-scale upper stage engine.

SpaceX was building Raptor anyway and asked for a much smaller percentage of the development cost from the government.  (It also has a potentially useful side effect of giving the government some insight into the development if SpaceX wants to get a Raptor based vehicle certified later).

I cross-edited you for the potential other uses of Raptor, but that's right.  Overall, AR1 was quoted at $804 million total and Raptor was quoted at $184 million total.

Offline woods170

  • IRAS fan
  • Senior Member
  • *****
  • Posts: 7338
  • IRAS fan
  • The Netherlands
  • Liked: 3057
  • Likes Given: 899
Re: Why is space so expensive
« Reply #50 on: 09/11/2017 06:08 AM »
You do know that SpaceX has acknowledge that it spent, or is spending, a cool $1 billion just to develop first stage recovery?  Given that number, it seems likely that Falcon 9 itself has and is costing multiple billions of dollars to develop.  Don't kid yourself about what this stuff costs.  SpaceX may have found a more cost-efficient path to develop its big rockets than others, but it still isn't cheap!

 - Ed Kyle
Falcon 9 v1.1 was all about first stage landing.
Falcon 9 v1.0 was all-new and cost roughly $800 million to develop-to-first-flight. The $1 billion mentioned for first stage landing constitutes the bulk of v1.1 development cost, including a major upgrade of the second stage.
The total cost of developing two major versions of Falcon 9 (v1.0 and v1.1-and-subsequent versions) is roughly $2.5 billion.
That's about the same as it cost to develop Atlas 5 but Atlas 5 didn't go thru two completely different versions, doesn't have first stage landing capabilities, uses an "off-the-shelf" main-stage engine from Russia as well as an "off-the-shelf" engine for it's upper stage. Atlas 5 had little development cost for RD-180 given that RD-180 was initially developed for Atlas III and Atlas V inherited RD-180 pretty much un-altered. The remaining development cost of RD-180 for Atlas V revolved around integration with the Atlas V first stage.
SpaceX however developed two versions of Merlin (1C and 1D) for Falcon 9 from scratch.

Like it or not Ed but a comparable amount of R&D investment resulted in a lot more Falcon 9 functionality than Atlas V functionality.
« Last Edit: 09/11/2017 12:45 PM by woods170 »

Online john smith 19

  • Senior Member
  • *****
  • Posts: 6159
  • Everyplaceelse
  • Liked: 814
  • Likes Given: 5105
Re: Why is space so expensive
« Reply #51 on: 09/11/2017 06:43 AM »
The only way to compete with constellations are with a overwhelming technology on orbit approach that doesn't care about long term reliability (because you keep on shoveling sats when needed), but does care about maximum yield (aggregate bandwidth) from the constellation, where you assume sat/launch costs are cheap/disposable. If you get them wrong it doesn't matter because the next salvo does better and eventually you win.
But keep in mind the underlying assumptions.

1)There is an unmet demand for the services that the constellation supplies and the way in which it supplies it.
It's possible to supply internet access from a comm sat in GEO, but unless you have very special needs it's the connection of very last resort for most people.   :(

2)That demand is growing at a rate that will justify that growth in bandwidth.

If those are correct you're right. I think satellite delivered services can be done better (both in bandwidth and costs) but as usual there's an army of Satan's little helpers lurking in the details.



"Solids are a branch of fireworks, not rocketry. :-) :-) ", Henry Spencer 1/28/11  Averse to bold? You must be in marketing."It's all in the sequencing" K. Mattingly.  STS-Keeping most of the stakeholders happy most of the time.

Online AncientU

  • Senior Member
  • *****
  • Posts: 4958
  • Liked: 2963
  • Likes Given: 4182
Re: Why is space so expensive
« Reply #52 on: 09/11/2017 10:57 AM »
The only way to compete with constellations are with a overwhelming technology on orbit approach that doesn't care about long term reliability (because you keep on shoveling sats when needed), but does care about maximum yield (aggregate bandwidth) from the constellation, where you assume sat/launch costs are cheap/disposable. If you get them wrong it doesn't matter because the next salvo does better and eventually you win.
But keep in mind the underlying assumptions.

1)There is an unmet demand for the services that the constellation supplies and the way in which it supplies it.
It's possible to supply internet access from a comm sat in GEO, but unless you have very special needs it's the connection of very last resort for most people.   :(

2)That demand is growing at a rate that will justify that growth in bandwidth.

If those are correct you're right. I think satellite delivered services can be done better (both in bandwidth and costs) but as usual there's an army of Satan's little helpers lurking in the details.

Serving a limited internet market from limited GEO slots will always be expensive.  That's the way it has been done, not the way it is planned to be done.

Internet demand is growing much faster than any space-based capacity could hope to cover.
Old, expensive, slow internet from GEO is on its way out...
"If we shared everything [we are working on] people would think we are insane!"
-- SpaceX friend of mlindner

Offline rayleighscatter

  • Full Member
  • ****
  • Posts: 942
  • Maryland
  • Liked: 428
  • Likes Given: 194
Re: Why is space so expensive
« Reply #53 on: 09/11/2017 02:06 PM »

Internet demand is growing much faster than any space-based capacity could hope to cover.
Old, expensive, slow internet from GEO is on its way out...
To be replaced with new, expensive, slow internet from LEO!

Offline Nibb31

  • Full Member
  • **
  • Posts: 231
  • France
  • Liked: 139
  • Likes Given: 8
Re: Why is space so expensive
« Reply #54 on: 09/11/2017 02:40 PM »
But keep in mind the underlying assumptions.

1)There is an unmet demand for the services that the constellation supplies and the way in which it supplies it.
It's possible to supply internet access from a comm sat in GEO, but unless you have very special needs it's the connection of very last resort for most people.   :(

2)That demand is growing at a rate that will justify that growth in bandwidth.

If those are correct you're right. I think satellite delivered services can be done better (both in bandwidth and costs) but as usual there's an army of Satan's little helpers lurking in the details.

Both of these assumptions are risky at this point.

If they are priced competitively, there is little doubt that LEO constellations will replace current satellite internet services for users such as:
- homes in remote locations
- ships and aircraft
- satphones

However, these are relatively small niche markets, which is proven by the fact that the large internet/mobile operators have pretty much ignored those markets. These are not the trillion dollar markets that will allow you build satellites and fund Mars colonies.

The largest part of the telecom market is the one that is covered by the established telecom industry. In most countries, national telecom companies are among the largest employers and are therefore extremely influential.

That industry is not going to sit and wait for SpaceX or OneWeb to steal the market from under their feet. They will do everything they can to remain competitive on land-based wireless and cable. They have a huge established infrastructure, are capable of very aggressive pricing, and will be introducing 5G networks by the time these constellations start to go online.

Remember also that the constellation operators will need to address  a global market, where they will be competing against domestic operators, who will often have strong political support.

I think it will be extremely difficult for SpaceX or Oneweb to penetrate the traditional ISP and mobile carrier markets. Those operators will do everything they can to contain constellation comms in their niche markets.

Online AncientU

  • Senior Member
  • *****
  • Posts: 4958
  • Liked: 2963
  • Likes Given: 4182
Re: Why is space so expensive
« Reply #55 on: 09/11/2017 04:38 PM »

Internet demand is growing much faster than any space-based capacity could hope to cover.
Old, expensive, slow internet from GEO is on its way out...
To be replaced with new, expensive, slow internet from LEO!

Then they'll be wasting billions on building and lofting the constellations...
(or you could be wrong)

At least in a free market, crappy solutions die a natural death.
« Last Edit: 09/11/2017 04:39 PM by AncientU »
"If we shared everything [we are working on] people would think we are insane!"
-- SpaceX friend of mlindner

Offline rayleighscatter

  • Full Member
  • ****
  • Posts: 942
  • Maryland
  • Liked: 428
  • Likes Given: 194
Re: Why is space so expensive
« Reply #56 on: 09/11/2017 05:13 PM »

Internet demand is growing much faster than any space-based capacity could hope to cover.
Old, expensive, slow internet from GEO is on its way out...
To be replaced with new, expensive, slow internet from LEO!
Wrong.

Faster, more reliable, greater volume, <130msec latency, ... eventually lower cost than ground internet. (If the correct investments/deployments/buyouts are made, you could get <80msec latency continental service on the ground.)

Long term expect surface networks to optimize latency (for trading). (Originally, believe it or not, Enron (yes that one) wanted to do an optimized latency internet service, close to speed of light limited, using MEMS/other "cut through" switching - was too early.)

Space routed core/peering via constellations will have the fastest to deploy latest improvements , moving to a 14-15 month cycle, ahead of consumer devices, becoming a Cisco+service provider vertical industry, until the technology slows down (if ever). Perhaps direct to consumer sometime, but that part's cloudy at the moment, as it is, the base station/metro will become the dominate "edge", less priced by location due to all sky hand off.

The internet is eating telco/cable/all comm services slowly. Hugely profitable industries are the chief focus of internet disruption. You'll end up with a $10T+ global service provider market - it's the one to own.

Right, all true of GEO as well. New names, same promises.

Offline meberbs

  • Full Member
  • ****
  • Posts: 1373
  • Liked: 1250
  • Likes Given: 317
Re: Why is space so expensive
« Reply #57 on: 09/11/2017 05:34 PM »

Internet demand is growing much faster than any space-based capacity could hope to cover.
Old, expensive, slow internet from GEO is on its way out...
To be replaced with new, expensive, slow internet from LEO!
Wrong.

Faster, more reliable, greater volume, <130msec latency, ... eventually lower cost than ground internet. (If the correct investments/deployments/buyouts are made, you could get <80msec latency continental service on the ground.)

Long term expect surface networks to optimize latency (for trading). (Originally, believe it or not, Enron (yes that one) wanted to do an optimized latency internet service, close to speed of light limited, using MEMS/other "cut through" switching - was too early.)

Space routed core/peering via constellations will have the fastest to deploy latest improvements , moving to a 14-15 month cycle, ahead of consumer devices, becoming a Cisco+service provider vertical industry, until the technology slows down (if ever). Perhaps direct to consumer sometime, but that part's cloudy at the moment, as it is, the base station/metro will become the dominate "edge", less priced by location due to all sky hand off.

The internet is eating telco/cable/all comm services slowly. Hugely profitable industries are the chief focus of internet disruption. You'll end up with a $10T+ global service provider market - it's the one to own.

Right, all true of GEO as well. New names, same promises.
Which of those statements do you think are true for GEO? GEO being "faster, more reliable, or greater volume" than GEO doesn't even make sense. <130msec latency is literally against the laws of physics for GEO.

Rapid replacements and upgrades were never done, and as far as I know never planned for GEO.

Offline whitelancer64

Re: Why is space so expensive
« Reply #58 on: 09/11/2017 05:48 PM »
I skimmed the thread and I didn't see anyone talk about this:

A big part of the reason that (aero)space is so expensive comes down to one word: traceability. From raw materials to completed assembly, all parts and all work done on a rocket must be traceable. That means lots and lots and lots of documentation that must be stored for decades so that if a part fails, then where the materials the part were made came from, who worked on it, what processes it underwent, what tests were performed on it, etc. can be looked at.

So you can have a $0.20 bolt that is used to put together 2x4s on a construction site, but if you put it in a rocket it winds up costing $25. What you are paying for is all the records associated with that part. What grade the raw materials were, and when, where, and how it was cast, who inspected it and when, what strength tolerances it was built and tested to, who shipped it, who received it, who inspected the part to ensure it wasn't damaged in transit, who installed it and what were the installation procedures, how was it confirmed that the installation was done correctly, and so on.

This is how it is possible to trace faults that happen on a probe tens of millions of miles away, to not only determine that X part failed, but how and why it failed, and that enables changes to be made to the way the part is made, or how it is installed, or whatever part of the process actually failed, to prevent a similar failure from happening in the future.

Traceability is a compounding factor on the price tag associated with low-quantity production for parts that go on space probes. Low volume of manufacture and one-off assembly, testing, calibration procedures, etc.
"One bit of advice: it is important to view knowledge as sort of a semantic tree -- make sure you understand the fundamental principles, ie the trunk and big branches, before you get into the leaves/details or there is nothing for them to hang on to." - Elon Musk
"There are lies, damned lies, and launch schedules." - Larry J

Online Space Ghost 1962

  • Senior Member
  • *****
  • Posts: 2474
  • Whatcha gonna do when the Ghost zaps you?
  • Liked: 2375
  • Likes Given: 1736
Re: Why is space so expensive
« Reply #59 on: 09/11/2017 06:05 PM »
Correct. And more. You need to know practically what went on in the minds of designers too, and test outcomes.

Am working with a startup on 100x reduction of this. It is a very hard problem, but the benefit is that you can greatly accelerate space payload/applications. It has take close to a decade. Has been in to DFJ already.

Tags: