Author Topic: How frequently does a RLV need to fly in order to amortize its costs?  (Read 14736 times)

Offline Pipcard

  • Full Member
  • ****
  • Posts: 622
  • Liked: 275
  • Likes Given: 130
"Traditionally, you need at least 8 launches per year to justify partial (first stage) reuse and at least 40 launches per year to justify full reuse." -  Robotbeat

I keep seeing claims like these or these and wonder, what study (or studies) came to this conclusion, and how did they do it? Of course, the reason for this flight rate is to amortize the costs of development and maintenance overhead, but to what extent?

Elon Musk also claims that "At a low flight rate, the improvement [in launch cost] is still probably around 50 percent," although the exact rate isn't specified.
« Last Edit: 06/28/2015 06:14 pm by Pipcard »

Offline guckyfan

  • Senior Member
  • *****
  • Posts: 7438
  • Germany
  • Liked: 2332
  • Likes Given: 2891
"Traditionally" is the operative word here. Falcon 9 is not a traditional RLV. While it would need a high number of flights to get to lowest cost and still maintain revenue flow for SpaceX it works at any number of reflights, starting with one. It would just lower the cost per flight less than with many flights.

Offline pippin

  • Regular
  • Senior Member
  • *****
  • Posts: 2575
  • Liked: 312
  • Likes Given: 45
Has nothing to do with traditional or not. There is an amount of fixed cost in your production chain and there is an amount of variable cost towards recovery and reflight. Unless you know these you can't determine the flight rate required to make reuse cheaper than building a new core but the flight rate will definitely be higher than one flight a year.

Offline john smith 19

  • Senior Member
  • *****
  • Posts: 10351
  • Everyplaceelse
  • Liked: 2431
  • Likes Given: 13606
Has nothing to do with traditional or not. There is an amount of fixed cost in your production chain and there is an amount of variable cost towards recovery and reflight. Unless you know these you can't determine the flight rate required to make reuse cheaper than building a new core but the flight rate will definitely be higher than one flight a year.
There is also the implicit assumption that the company that builds the RLV will operate the entire fleet.

So imagine the cost implications of Boeing running an airline and only building aircraft for them and only getting revue from their own flights. [EDIT versus the number of aircraft of any Boeing design they actually sell to all air lines ]
« Last Edit: 06/29/2015 02:34 pm by john smith 19 »
MCT ITS BFR SS. The worlds first Methane fueled FFSC engined CFRP SS structure A380 sized aerospaceplane tail sitter capable of Earth & Mars atmospheric flight.First flight to Mars by end of 2022 TBC. T&C apply. Trust nothing. Run your own #s "Extraordinary claims require extraordinary proof" R. Simberg."Competitve" means cheaper ¬cheap SCramjet proposed 1956. First +ve thrust 2004. US R&D spend to date > $10Bn. #deployed designs. Zero.

Offline Ludus

  • Full Member
  • ****
  • Posts: 1744
  • Liked: 1255
  • Likes Given: 1017
If it's like the shuttle and is ONLY a reusable launch vehicle (RLV) the answer is different than if it's like F9R and can also be expendable. The shuttle style vehicle isn't produced in quantity, it's designed to be reused many times and it's got a unit cost MUCH higher than comparable launch capacity on expendable rockets. It has to fly enough to offset that difference when all reuse and fixed costs are accounted for in order to amortize it's cost. The issue only has meaning relative to what launch would have cost otherwise.

The F9R is profitable with a single use at current prices so it doesn't need to fly more than once to amortize it's cost. Recovery and reuse doesn't matter. if SpaceX wanted to lower launch prices for reuse they could, but it would cannibalize some business. Since SpaceX is currently the low cost provider they would be undercutting themselves.

The marginal cost for reflying a F9R could be extremely low (gas and go) even an order of magnitude lower than for expending a new rocket and still not make the case for lowering the customer price of a launch at all. SpaceX might take advantage of lower marginal costs with internal projects like the Internet constellation without changing prices.

Offline john smith 19

  • Senior Member
  • *****
  • Posts: 10351
  • Everyplaceelse
  • Liked: 2431
  • Likes Given: 13606
If it's like the shuttle and is ONLY a reusable launch vehicle (RLV) the answer is different than if it's like F9R and can also be expendable.
The F9 is an ELV that  SX are trying to turn into a semi reusable LV. It could be be argued that is no more than the same level of reusability as the Shuttle.
Quote
The shuttle style vehicle isn't produced in quantity, it's designed to be reused many times and it's got a unit cost MUCH higher than comparable launch capacity on expendable rockets.
the argument is an RLV over it's lifetime can put many times the payload of the biggest ELV into orbit.  The question then becomes how many giant monolithic payloads exist to support that size of ELV. The answer is not many.
Quote
It has to fly enough to offset that difference when all reuse and fixed costs are accounted for in order to amortize it's cost. The issue only has meaning relative to what launch would have cost otherwise.
Again this assumes the builder is the operator.  There is no other transport mode where this is the case. Shipyards don't run cruise lines. Truck makers don't run haulage companies.

This suggests there is something quite wrong with how the launcher and launch services market operates.
Quote
The F9R is profitable with a single use at current prices so it doesn't need to fly more than once to amortize it's cost. Recovery and reuse doesn't matter. if SpaceX wanted to lower launch prices for reuse they could, but it would cannibalize some business. Since SpaceX is currently the low cost provider they would be undercutting themselves.
The fully reusable F9 has been off the table since September 25th 2014 at the latest.
Quote
The marginal cost for reflying a F9R could be extremely low (gas and go) even an order of magnitude lower than for expending a new rocket and still not make the case for lowering the customer price of a launch at all. SpaceX might take advantage of lower marginal costs with internal projects like the Internet constellation without changing prices.
If it existed to begin with.

Which it does not.

So I guess you're saying that the F9, being an ELV, just needs to fly once to cover its costs.
Just like every other ELV.
MCT ITS BFR SS. The worlds first Methane fueled FFSC engined CFRP SS structure A380 sized aerospaceplane tail sitter capable of Earth & Mars atmospheric flight.First flight to Mars by end of 2022 TBC. T&C apply. Trust nothing. Run your own #s "Extraordinary claims require extraordinary proof" R. Simberg."Competitve" means cheaper ¬cheap SCramjet proposed 1956. First +ve thrust 2004. US R&D spend to date > $10Bn. #deployed designs. Zero.

Offline ClaytonBirchenough

  • Full Member
  • ****
  • Posts: 777
  • ~ 1 AU
  • Liked: 34
  • Likes Given: 348
We'll have to see.

Seriously though. The brightest minds (at the time) engineered the STS. It offered lower cost and a higher flight rate. The program did not deliver on its reusability promise and that affected its flight rate which then affected its cost.

Numerous programs, studies, and companies have produced viable reusable vehicles on paper and with prototypes, but they often fail on practicality.

It seems (with regards to SpaceX's F9R) the first question we should be asking is how much refurbishment (cost wise) after each flight F9R's first stage needs for reuse. Then SpaceX figures out in reality what it actually costs.

Not sure if what I said just made sense, but anyone that has worked in or studied the LV industry should know that studies and paper rockets are often "optimistic" and don't follow through with their "promises". Any speculation based on these unproven studies and paper rockets becomes that more unpredictable and inaccurate.
Clayton Birchenough

Offline SoulWager

  • Full Member
  • *
  • Posts: 177
  • Liked: 59
  • Likes Given: 11
Depends on the elasticity of demand, on the marginal cost of a new core, on the marginal cost of refurbishment, on fixed costs in both production and refurbishment, and on reliability.

Elasticity of demand is one big factor I think ULA failed to properly account for, so that's why they expect a fully reusable vehicle to need more flights to break even. If the reused flights get sold in addition to the existing manifest, you need far fewer flights per core to make the reuse case profitable.

Offline guckyfan

  • Senior Member
  • *****
  • Posts: 7438
  • Germany
  • Liked: 2332
  • Likes Given: 2891
So I guess you're saying that the F9, being an ELV, just needs to fly once to cover its costs.
Just like every other ELV.

Yes, just with the bonus that its first stage can fly again. :)

Offline guckyfan

  • Senior Member
  • *****
  • Posts: 7438
  • Germany
  • Liked: 2332
  • Likes Given: 2891
Depends on the elasticity of demand,

That's the brilliant economics in the SpaceX reuse concept. It works at any point of the elasticity curve. It can lower prices if more flights occur, it is very competetive if not.

Offline pippin

  • Regular
  • Senior Member
  • *****
  • Posts: 2575
  • Liked: 312
  • Likes Given: 45
Has nothing to do with traditional or not. There is an amount of fixed cost in your production chain and there is an amount of variable cost towards recovery and reflight. Unless you know these you can't determine the flight rate required to make reuse cheaper than building a new core but the flight rate will definitely be higher than one flight a year.
There is also the implicit assumption that the company that builds the RLV will operate the entire fleet.

So imagine the cost implications of Boeing running an airline and only building aircraft for them and only getting revue from their own flights. [EDIT versus the number of aircraft of any Boeing design they actually sell to all air lines ]

I can't imagine those because they would completely depend on how competent Boeing is at managing an airline, something we all don't know. There are no implicit cost implications of vertical integration, only business model implications.
« Last Edit: 07/06/2015 12:02 pm by pippin »

Offline pippin

  • Regular
  • Senior Member
  • *****
  • Posts: 2575
  • Liked: 312
  • Likes Given: 45
Depends on the elasticity of demand,

That's the brilliant economics in the SpaceX reuse concept. It works at any point of the elasticity curve. It can lower prices if more flights occur, it is very competetive if not.

This is wrong. Their fixed costs don't go away. They lower the price and then can no longer pay their staff PLUS they have to pay for whatever maintenance the recovery requires.
Nothing gained from the recovered core if demand isn't high enough to cover the fixed cost.
« Last Edit: 07/06/2015 12:05 pm by pippin »

Offline JamesH

  • Full Member
  • ****
  • Posts: 525
  • United Kingdom
  • Liked: 284
  • Likes Given: 7
But they presumably have wiggle room to reduce the costs before they hit the 'staff' costs issues, and with the EELV versions, they have no recovery costs.

But for the moment, even with re-usability, they probably need to keep prices up to pay for Mars, but conversely, drop them enough to encourage more customers. Standard business balancing act.

Offline Wigles

  • Member
  • Posts: 52
  • Liked: 12
  • Likes Given: 5
Depends on the elasticity of demand,

That's the brilliant economics in the SpaceX reuse concept. It works at any point of the elasticity curve. It can lower prices if more flights occur, it is very competetive if not.

This is wrong. Their fixed costs don't go away. They lower the price and then can no longer pay their staff PLUS they have to pay for whatever maintenance the recovery requires.
Nothing gained from the recovered core if demand isn't high enough to cover the fixed cost.

The price to reuse can't be higher than the expendable cost (or reuse is an economic folly), and their sell price can't be lower than their fixed costs; this defines the upper and lower bounds. However, they are apparently profitable at the expendable end of the spectrum so as long as reuse costs are less than new build costs then increasing reuse should lead to lower costs.

The question of frequency shouldn't come into it unless the storage costs (or other time-based costs) for a non-flying vehicle are a significant cost driver (which may be the case, not a professional reusable rocket accountant).

This cost reduction is going to be a diminishing returns scenario trending to a little higher than their fixed costs, and there will come a point (number of reuses) whereby the cost of refurbishment actually starts to increase. That will define the lowest cost point.

The question of how many times a reusable rocket needs to fly to amortize its costs is backwards. I would want to understand how many times I could reuse a rocket (or stage) and what the cost per reuse is and how that trends with number of reuses. That will then allow me to define my pricing such that I can spread out the unit costs among enough launches and still make a profit overall.

Offline Wigles

  • Member
  • Posts: 52
  • Liked: 12
  • Likes Given: 5
But for the moment, even with re-usability, they probably need to keep prices up to pay for Mars, but conversely, drop them enough to encourage more customers. Standard business balancing act.

Agreed, if they crack reuse and the reuse costs are not significant compared to the production costs then they can easily afford to undercut their competition to steal market share while fostering new growth while also retaining a healthy profit margin for other activities Elon might have his eye on.

Offline pippin

  • Regular
  • Senior Member
  • *****
  • Posts: 2575
  • Liked: 312
  • Likes Given: 45
But they presumably have wiggle room to reduce the costs before they hit the 'staff' costs issues, and with the EELV versions, they have no recovery costs.
Huh? There is nothing in this world, that has "no costs". Everything comes as a price and of course they have recovery costs. A lot of which are fixed costs again which makes the economies of the whole thing even more susceptible to changes in demand.

Offline pippin

  • Regular
  • Senior Member
  • *****
  • Posts: 2575
  • Liked: 312
  • Likes Given: 45
The price to reuse can't be higher than the expendable cost (or reuse is an economic folly)
Yea, but who says it isn't? SpaceX for sure doesn't know, yet. They invest to find out.

Quote
and their sell price can't be lower than their fixed costs
???
Of course they can. That's what every growing company does. I bet SpaceX has yet to sell the first flight that covers their fixed costs. Right now they are still scaling up so they are far from covering their fixed costs with the ongoing business.

Quote
However, they are apparently profitable at the expendable end of the spectrum
Who says they are profitable? SpaceX doesn't, SpaceX says they are cash-flow-positive which is something entirely different. I would be very, very surprised if they were profitable.

Profitability is also an accounting figure that is primarily of interest for capital investors, not actual operational business.

Quote
so as long as reuse costs are less than new build costs then increasing reuse should lead to lower costs.
This wouldn't even be true if they were profitable.
We don't know SpaceX' cost structure, but they are very vertically integrated.
If their fixed cost (factory, people, capital cost, launch pads,....) makes a high part of their overall cost (I wouldn't be surprised, if that's well beyond 70-80% in SpaceX's case, that raw AlLi and copper they buy isn't really a massive cost driver) then there is very little room for savings. If not building a core only saves you 25% of the overall flight costs then everything you need for the reuse, including fixed infrastructure, personnel, capex,... needs to be below 25% of the flight price to have a working business case. That changes only at the moment where you can increase the flight rate beyond the capacity of your infrastructure (factory), at that point you start to profit massively because you don't need additional investment. Those incremental flights will then come at very low incremental cost and will be profitable independent of the state of the rest of your business.
But as far as your flight rate is too low to pay for all the capacity you built it will be very hard to find any business case for reuse, no matter how cheap it is.

So it all depends on the flight rate and we don't know that.

Offline guckyfan

  • Senior Member
  • *****
  • Posts: 7438
  • Germany
  • Liked: 2332
  • Likes Given: 2891
Depends on the elasticity of demand,

That's the brilliant economics in the SpaceX reuse concept. It works at any point of the elasticity curve. It can lower prices if more flights occur, it is very competetive if not.

This is wrong. Their fixed costs don't go away. They lower the price and then can no longer pay their staff PLUS they have to pay for whatever maintenance the recovery requires.
Nothing gained from the recovered core if demand isn't high enough to cover the fixed cost.

If you read my post this is exactly what I wrote. As an ELV they are very competetive right now. They will be more competetive when they can refly but the trade how much they can reduce prices depends on the number of flights they can spread the fixed cost over. So very limited price reductions when the number of flights does not increase, a lot of price reduction when it increases a lot. But unlike a vehicle like SpaceShuttle or Skylon which have skyhigh fixed cost they can compete at any number of flights, high or low.

Offline guckyfan

  • Senior Member
  • *****
  • Posts: 7438
  • Germany
  • Liked: 2332
  • Likes Given: 2891
Who says they are profitable? SpaceX doesn't, SpaceX says they are cash-flow-positive which is something entirely different. I would be very, very surprised if they were profitable.

There are statements that indicate they need about 15 launches a year to be profitable once NASA payments for development are gone.

Edit: fixed quote
« Last Edit: 07/06/2015 01:04 pm by guckyfan »

Offline pippin

  • Regular
  • Senior Member
  • *****
  • Posts: 2575
  • Liked: 312
  • Likes Given: 45
As an ELV they are very competetive right now.
Because they are pricing their product competitively. We don't know their cost structure so we don't know whether they earn or lose money with that.

Quote
They will be more competetive when they can refly
Again: we don't know that. I bet neither does SpaceX. If it means they need all that infrastructure, landing pad, legs, drone ships, range assets, reduced payload yet only successfully recover every 3rd core or so they will not. If they can dismantle most of that in the long run and essentially successfully fly back 90% of the cores at high flight rate it's a no-brainer. Reality will likely be somewhere in between.

Quote
But unlike a vehicle like SpaceShuttle or Skylon which have skyhigh fixed cost they can compete at any number of flights, high or low.
That's a fallacy. They will be able to compete at a lower flight rate, that's for sure. Any? No way, there's fixed infrastructure they have to keep around for recovery and I bet that's more expensive than raw metal.

Offline pippin

  • Regular
  • Senior Member
  • *****
  • Posts: 2575
  • Liked: 312
  • Likes Given: 45
There are statements that indicate they need about 15 launches a year to be profitable once NASA payments for development are gone.

Source? Never, really, never, rely on hearsay for financial information. There's so much misunderstanding going on and it's so easy to make people misunderstand things you are saying. What does "need" in your example mean? To cover their fixed costs as they are now? To not have to shut down the whole company? To generate an operating profit? To create a return on the capital investment going into them? To fund their growth plans? Massive difference between these figures.
« Last Edit: 07/06/2015 01:14 pm by pippin »

Offline guckyfan

  • Senior Member
  • *****
  • Posts: 7438
  • Germany
  • Liked: 2332
  • Likes Given: 2891
There are statements that indicate they need about 15 launches a year to be profitable once NASA payments for development are gone.

Source? Never, really, never, rely on hearsay for financial information. There's so much misunderstanding going on and it's so easy to make people misunderstand things you are saying. What does "need" in your example mean? To cover their fixed costs as they are now? To not have to shut down the whole company? To generate an operating profit? To create a return on the capital investment going into them? To fund their growth plans? Massive difference between these figures.

I have problems with providing the source. But that statement was made not too long ago. If I recall correctly he said 12 flights a year but as SpaceX staff has increased since then it was extrapolated to 15/16 flights. I assume he or the investors do not care much about return on investment. That comes with increasing value of the company. As many of that staff is involved in development it would include the ongoing development efforts for the next generation rocket, Raptor and MCT. Probably not at present level for the satellite network. And not for building the factory and launch pad that requires. That is a major effort which will need additional funding beyond the 1 Billion $ Google put into the company recently.

Offline pippin

  • Regular
  • Senior Member
  • *****
  • Posts: 2575
  • Liked: 312
  • Likes Given: 45
I assume he or the investors do not care much about return on investment.
Depends on whether it's equity or debt. Valuation of the company depends on prospect of long-term returns, so investors usually care a lot, they just care even more about growth potential. Because that means you might generate more return.

Quote
As many of that staff is involved in development it would include the ongoing development efforts for the next generation rocket, Raptor and MCT. Probably not at present level for the satellite network. And not for building the factory and launch pad that requires. That is a major effort which will need additional funding beyond the 1 Billion $ Google put into the company recently.
Exactly. Which is why  say it's difficult to assess.
To some it all up: while you all claim to know very well how SpaceX's business model works all I'm saying is: you don't know any of that. And I bet a lot of the things you are all taking for granted here are high on the "I'd really like to understand this better" list of Elon.

Everybody always seems to assume that developing a business is all about having that great scheme of things in mind and knowing 10 years in advance what will happen while in reality it's nothing like that. It's all about seeing an opportunity, seeing some things you might be able to do better than others and then trying and trying. Even when you are successful, you usually find that 75% of the things you try simply don't work out and then you go on and try something else. As long as the other 25% work well enough that's perfectly fine.

As an entrepreneur you don't assume you know everything and make no mistakes, you assume you will get the important things done even if you make mistakes and make sure to shed the ones that don't work soon enough. And you try a lot of things.
« Last Edit: 07/06/2015 01:39 pm by pippin »

Offline john smith 19

  • Senior Member
  • *****
  • Posts: 10351
  • Everyplaceelse
  • Liked: 2431
  • Likes Given: 13606
We'll have to see.

Seriously though. The brightest minds (at the time) engineered the STS. It offered lower cost and a higher flight rate.
Technically the lower cost comes from the higher flight rate.
Quote
The program did not deliver on its reusability promise and that affected its flight rate which then affected its cost.
You've missed out the elephant that was in that room.

The absurd fundingr restriction placed on the programme. From then all all cost goals went out the window in order to get something built.
Quote
Numerous programs, studies, and companies have produced viable reusable vehicles on paper and with prototypes, but they often fail on practicality.
The paper parts correct.

Which ones were you thinking of that went to prototype?
Quote
It seems (with regards to SpaceX's F9R) the first question we should be asking is how much refurbishment (cost wise) after each flight F9R's first stage needs for reuse. Then SpaceX figures out in reality what it actually costs.
Sounds like  you need a cost model you can change the numbers on to see what converges.
Quote
Not sure if what I said just made sense, but anyone that has worked in or studied the LV industry should know that studies and paper rockets are often "optimistic" and don't follow through with their "promises".
Indeed. In the case of the X30 some of the values for the actual physical constants were wrong.

A fact the USAF only discovered about a $Bn into the programme.  :(
Quote
Any speculation based on these unproven studies and paper rockets becomes that more unpredictable and inaccurate.
Yes, the more hardware, and the more accurate hardware that gets built the more likely it will do it claims.
MCT ITS BFR SS. The worlds first Methane fueled FFSC engined CFRP SS structure A380 sized aerospaceplane tail sitter capable of Earth & Mars atmospheric flight.First flight to Mars by end of 2022 TBC. T&C apply. Trust nothing. Run your own #s "Extraordinary claims require extraordinary proof" R. Simberg."Competitve" means cheaper ¬cheap SCramjet proposed 1956. First +ve thrust 2004. US R&D spend to date > $10Bn. #deployed designs. Zero.

Offline john smith 19

  • Senior Member
  • *****
  • Posts: 10351
  • Everyplaceelse
  • Liked: 2431
  • Likes Given: 13606
So I guess you're saying that the F9, being an ELV, just needs to fly once to cover its costs.
Just like every other ELV.

Yes, just with the bonus that its first stage can fly again. :)
I'm looking forward to that day.

Depends on the elasticity of demand,

That's the brilliant economics in the SpaceX reuse concept. It works at any point of the elasticity curve. It can lower prices if more flights occur, it is very competetive if not.
In which case why do you offer reuse when you're already the lowest cost provider in the first place?

That's the paradox of companies operating a builder/operator model for an RLV, or in SX's case a semi reusable LV.

I can't imagine those because they would completely depend on how competent Boeing is at managing an airline, something we all don't know. There are no implicit cost implications of vertical integration, only business model implications.
According to SX every dollar spend with a subcontractor requires them to spend about $3-$5 in house.

That sounds like quite a large cost implication to me, if you're not on a government contractor where you know the taxpayer will pick up the tab.

In fact I'd like to extend my Boeing airline analogy further.

What if (in addition to running the airline) Boeing planes could operate only from Boeing owned airports?

So now every passenger directly carries the burden of a)The planes design team b)The planes build team c) The planes construction cost d)The airlines staff costs d)The airports construction cost e) The airports operating cost.

Do you doubt there will be some influence on the volume of air passenger travel and it's likely to be negative?

I will repeat no other transport mode operates in this way. If anything it's the "IBM Mainframe" model of the 1905's and 1960's.

I would suggest that as long as an LV architecture starts with a TSTO (or nSTO) ELV that is the way it will always operate due to the deep coupling of payload, vehicle and launch pad.  :(
« Last Edit: 07/06/2015 02:23 pm by john smith 19 »
MCT ITS BFR SS. The worlds first Methane fueled FFSC engined CFRP SS structure A380 sized aerospaceplane tail sitter capable of Earth & Mars atmospheric flight.First flight to Mars by end of 2022 TBC. T&C apply. Trust nothing. Run your own #s "Extraordinary claims require extraordinary proof" R. Simberg."Competitve" means cheaper ¬cheap SCramjet proposed 1956. First +ve thrust 2004. US R&D spend to date > $10Bn. #deployed designs. Zero.

Offline guckyfan

  • Senior Member
  • *****
  • Posts: 7438
  • Germany
  • Liked: 2332
  • Likes Given: 2891
Depends on the elasticity of demand,

That's the brilliant economics in the SpaceX reuse concept. It works at any point of the elasticity curve. It can lower prices if more flights occur, it is very competetive if not.
In which case why do you offer reuse when you're already the lowest cost provider in the first place?

Because Elon Musk is driven by the desire to advance spaceflight, not making profits. Of course he needs to make profits to invest towards his goal. He cannot fund it with profits from elsewhere, like Blue Origins Bezos.

That's the paradox of companies operating a builder/operator model for an RLV, or in SX's case a semi reusable LV.

Your fixation on separating building and operation is a source of constant amazement to me.

Offline pippin

  • Regular
  • Senior Member
  • *****
  • Posts: 2575
  • Liked: 312
  • Likes Given: 45
What if (in addition to running the airline) Boeing planes could operate only from Boeing owned airports?
Well, if it meant Boeing could rake in all the subsidies going into airport construction around the world that alone would be a massive push to their business models. Airports are money sinks and the cost is not carried by the airlines. Which, again, aren't profitable on average and who are _also_ massively subsidized in many places around the world.

No, the big advantage of separating the two business models of building and operating airliners is of course that you have competition on both sides so that you can wipe out the more incompetent ones and only leave the better ones as has happened numerous times instead of having all the airline operations go down with the inefficient airliner manufacturer or the other way around.

But that depends a lot on the level of competition you can achieve and the frequency of use.
Airline business as it is developed after WWII when lots of unused transport aircraft were available cheaply, to this day it's never really become a profitable sector.

Quote
I will repeat no other transport mode operates in this way.
Yea, and no transport sector is profitable on it's own, they all get their infrastructure subsidized. All of them.

Quote
If anything it's the "IBM Mainframe" model of the 1905's and 1960's.
Or the Google- and Facebook model of today.


Offline Coastal Ron

  • Senior Member
  • *****
  • Posts: 8862
  • I live... along the coast
  • Liked: 10199
  • Likes Given: 11934
In which case why do you offer reuse when you're already the lowest cost provider in the first place?

It depends on what your motivations are.  For Boeing and Lockheed Martin they did not see a business advantage in lowering the cost to access space, and it could be argued that they actually wanted to increase it when they merged.

However Elon Musk is not motivated purely by profit, as he has a personal goal of making humanity multi-planetary.  So while you need a profitable business in order to help make that goal a reality, it does open you up to changing the economics of how everything works.

To be fair, if there were far more competitors it is likely someone would have been as aggressive as Musk in pursuing reusability, but unfortunately there aren't.

Quote
That's the paradox of companies operating a builder/operator model for an RLV, or in SX's case a semi reusable LV.

We're not far enough down the road yet to understand if SpaceX will always be a builder/operator.  For Musk's personal goals it might be better that they turn into a builder only, and sell to operators - like the airline model.  But if that does happen it might take a while, so for now what we may be seeing as a builder/operator may not represent where the market will eventually go.

Quote
I would suggest that as long as an LV architecture starts with a TSTO (or nSTO) ELV that is the way it will always operate due to the deep coupling of payload, vehicle and launch pad.  :(

No doubt there are still major pieces missing to make a really nice commercial space transportation market.  We're still truly in the early days here, with only guesses as to what will happen in the future.
If we don't continuously lower the cost to access space, how are we ever going to afford to expand humanity out into space?

Offline john smith 19

  • Senior Member
  • *****
  • Posts: 10351
  • Everyplaceelse
  • Liked: 2431
  • Likes Given: 13606
Well, if it meant Boeing could rake in all the subsidies going into airport construction around the world that alone would be a massive push to their business models. Airports are money sinks and the cost is not carried by the airlines. Which, again, aren't profitable on average and who are _also_ massively subsidized in many places around the world.
Or national governments could say "You want into our market, you fund it," or indeed a range of other options, depending on wheather they had airline mfgs and/or operators of their own.
Quote
No, the big advantage of separating the two business models of building and operating airliners is of course that you have competition on both sides so that you can wipe out the more incompetent ones and only leave the better ones as has happened numerous times instead of having all the airline operations go down with the inefficient airliner manufacturer or the other way around.
Which kind of suggest the people who run airliner builders really well don't make good airline operators, and vice versa.

Actually that seems to work with all forms of transport except the launcher/launch services business.
Quote
Yea, and no transport sector is profitable on it's own, they all get their infrastructure subsidized. All of them.
The trouble is the subsidy in launch is either tied one type of LV (no such thing as a "universal pad") or range costs are sized on a "per vehicle" basis, so a small LV gets no benefit in lower costs for being small.

Or in anti trust speak the "barriers to entry" are very high.
Quote
Quote
If anything it's the "IBM Mainframe" model of the 1905's and 1960's.
Or the Google- and Facebook model of today.
Yes I'd agree with that.

Although I think Google and Facebook are much smarter at not looking like a monopoly.
MCT ITS BFR SS. The worlds first Methane fueled FFSC engined CFRP SS structure A380 sized aerospaceplane tail sitter capable of Earth & Mars atmospheric flight.First flight to Mars by end of 2022 TBC. T&C apply. Trust nothing. Run your own #s "Extraordinary claims require extraordinary proof" R. Simberg."Competitve" means cheaper ¬cheap SCramjet proposed 1956. First +ve thrust 2004. US R&D spend to date > $10Bn. #deployed designs. Zero.

Offline pippin

  • Regular
  • Senior Member
  • *****
  • Posts: 2575
  • Liked: 312
  • Likes Given: 45
Which kind of suggest the people who run airliner builders really well don't make good airline operators, and vice versa.
Know any who tried? All it means is that builders of airliners not necessarily make the best airline operators and vice versa. Or run lucky in both businesses.

Quote
Actually that seems to work with all forms of transport except the launcher/launch services business.
Ariane has been working quite well for quite a while.

Quote
Although I think Google and Facebook are much smarter at not looking like a monopoly.
Huh? How's that? Both are clearly being seen as monopolies in their respective business by pretty much everyone. The point here, though, is that they are also very much vertically integrated down to the point of developing and building their own hardware for their data centers.
« Last Edit: 07/06/2015 03:15 pm by pippin »

Offline Jim

  • Night Gator
  • Senior Member
  • *****
  • Posts: 37442
  • Cape Canaveral Spaceport
  • Liked: 21452
  • Likes Given: 428

Which kind of suggest the people who run airliner builders really well don't make good airline operators, and vice versa.

Not true.  See United Aircraft and Transport Corporation.

Offline Jim

  • Night Gator
  • Senior Member
  • *****
  • Posts: 37442
  • Cape Canaveral Spaceport
  • Liked: 21452
  • Likes Given: 428

The trouble is the in launch is either tied one type of LV (no such thing as a "universal pad") or range costs are sized on a "per vehicle" basis, so a small LV gets no benefit in lower costs for being small.


No.

A.  range services are independent of vehicle size.  A smaller vehicle doesn't use less range services.  They may use less base services which are cost reimbursable.  But items like telemetry, range safety, range clearance are basically the same cost no matter the size.
b.  it is also independent of with the pad. subsidy is based on the launch service and not the hardware used to launch the vehicle.

Offline ClaytonBirchenough

  • Full Member
  • ****
  • Posts: 777
  • ~ 1 AU
  • Liked: 34
  • Likes Given: 348
The question of frequency shouldn't come into it unless the storage costs (or other time-based costs) for a non-flying vehicle are a significant cost driver (which may be the case, not a professional reusable rocket accountant).

Frequency is a very key point in reducing costs per launch.

If I have 1 RLV and launch it once a year, I have to cover all overhead costs with that one launch! If I launch that RLV 100 times, I only have to cover .01 of my overhead for the year on each launch.

There are many things that need to be amortized on each launch. Development costs, labor, etc. And, in such a drastic exaggeration as I used above, you have your employees refurbish the RLV for reuse. This certainly doesn't take a year. And my bet is, you don't exactly just bring in your "refurbishers" on a part time basis. Economies of scale seem to be your friend in the rocket business. :)

I am by no means a professional reusable rocket accountant either, but from my perspective, it appears that facility and development costs are a large driver in LV prices.

(Correct me if I'm wrong :) )
Clayton Birchenough

Offline john smith 19

  • Senior Member
  • *****
  • Posts: 10351
  • Everyplaceelse
  • Liked: 2431
  • Likes Given: 13606
People have talked a lot about qualitative analysis, range costs, re use rates etc.

With this in mind here is a copy of my current cost model

The 1st tab is for the F9SR design, the next BFR, the third the MCT and the last Skylon.

To use the model put your preferred values in the blue bordered cells and the model shows the results. Cells with red dots have help text or store known or preferred constants for various parameters, in case you overwrite a value and decide you want to try again with the original.

[EDIT. A special note on units used on a cell. Read the relevant label. If you put 0.1 in a cell labelled "percent" that's 0.1%, not 10%. Like wise different costs are entered in $, 1000s or $m. The model tries to ask for entry in reasonable sized units to limit the number of zeros people have to enter if the whole thing was in $ only. I'll consider changing if people really feel the need to see everything to the $]

Certain design options are triggered putting a Y or N in the relevant labelled cells, for example if  you think the staff costs for refurbishment should be charged by number of hours needed on a launch (unrealistic unless the number of launches a year is very high and they are like cars on an assembly line at Ford) or on number of launches per  year

Both number of flights per year and for the overall life of the F9SR are adjustable.

This new version breaks out range usage costs as a separate item (but assumes the range already exists) and lets you set a fixed charge on top of costs rather than a percentage of costs as the final asking price, if you think SX will ride the "cost curve" down while keeping the same level of profit

As previously there is an "arbitrary cost" box (which you can switch off) for a number to cover all the stuff  you can't specifically think that a refurbishment will need, but you're sure it will need.

Error reports, suggestions for improvements and other items to break out as specific costs are all welcome.
« Last Edit: 07/07/2015 12:57 am by john smith 19 »
MCT ITS BFR SS. The worlds first Methane fueled FFSC engined CFRP SS structure A380 sized aerospaceplane tail sitter capable of Earth & Mars atmospheric flight.First flight to Mars by end of 2022 TBC. T&C apply. Trust nothing. Run your own #s "Extraordinary claims require extraordinary proof" R. Simberg."Competitve" means cheaper ¬cheap SCramjet proposed 1956. First +ve thrust 2004. US R&D spend to date > $10Bn. #deployed designs. Zero.

Offline pippin

  • Regular
  • Senior Member
  • *****
  • Posts: 2575
  • Liked: 312
  • Likes Given: 45
Sorry, but your cost model assumes a fixed cost structure with only variable costs and no changes to cost structure with utilization, this is completely unrealistic and does costing totally backwards. You make the biggest mistake right at the beginning by assuming something like a fixed "profit" per flight. They might calculate with that at a given target flight rate but it's not a fixed value. One more flight and it goes up, one less flight and it goes down. Even if you know these figures (the assumptions at a given target flight rate) the results would only be valid for a static model with exactly this rate and no changes and can't be used to calculate anything else.

Start with assumptions about manufacturing costs, capex, capacity, material cost, variable costs per launch and so on and build on that.

Your model is something you might be able to use to deduct SpaceX's cost structures if you _knew_ all these other figures (cost per launch, overall profit, both of the for a certain number of flights per year, utilization, investment,...) which would then allow you to build on that but since you know none of them except for some vaguely defined gross figures you can't use it to deduct anything.
« Last Edit: 07/07/2015 01:22 am by pippin »

Offline pippin

  • Regular
  • Senior Member
  • *****
  • Posts: 2575
  • Liked: 312
  • Likes Given: 45
If you want to build a  cost model you essentially have two options:

1. build a real cost structure. Here you have to know things like development costs (over several years), investments, depreciation for the investments (or usage times), capacity, labor cost, capital cost,.... then you have to create a model for the fixed costs for a given capacity range and a change to these fixed costs if you expand/contract that capacity and then variable (marginal) costs per flight. Then you can either calculate a price or a profit/loss (if you have the other). Or you can calculate the sensitivity of price/profit to changes to the flight rate/number of flights. Within your capacity constraints.

2. build a cash-flow model. Here you can assume how much cash inflow and outflow they have per year and derive at what rates they can keep operating with a given amount of investment they have ("cash burn rate") or make ("positive cash flow").
With this you can not, though, deduct things like "profits" of any kind, cost structures or costs per flight or per unit and prices (which are generally based as much on market conditions as on actual costs), what you _can_ deduct is the impact on the cash use/accumulation from a changing flight rate. That's probably the best information you can derive from the little data that's really known about SpaceX's business.

Familiarize yourself with concepts like fixed cost and marginal costs, otherwise this whole exercise leads nowhere.

Also, please make sure you understand terms like "profit". "Profit" is an accounting figure you create as a result of what you did over a year. You might then deduct figures like "profit per flight" simply by dividing the profit by the number of flights but that doesn't mean you made that amount of profit per flight. Could be you just broke even with the last flight and your profit actually doubles by having an additional flight. Or your profit could have come from other sources like services and your profit actually lowers if you add another flight. It's meaningless.

What you _can_ use are figures like net cash flow per flight (difference between price and marginal cost) but that's what you then have to use to pay all your fixed costs which will be the lion's share of your overall costs. And even "fixed" costs change when you get out of your defined capacity range. Costing isn't simple if you don't have raw cost structure data.
« Last Edit: 07/07/2015 01:36 am by pippin »

Offline john smith 19

  • Senior Member
  • *****
  • Posts: 10351
  • Everyplaceelse
  • Liked: 2431
  • Likes Given: 13606
Sorry, but your cost model assumes a fixed cost structure with only variable costs and no changes to cost structure with utilization, this is completely unrealistic and does costing totally backwards.
It's actually a bit worse that that.  :)

I don't care what those numbers are.

The model is about what happens to the price (and possible profit, IE what they make above what they directly have to put in to make the LV and later refurb it) of an SX launch if 1st stage reuse happens (which still has not happened).

If SX are not making a profit at this level they will run out of money and go bankrupt. If they are they won't. If they are not making enough profit then their prices will likely rise. My interest is in lowering the price of launch and what that takes.

Will putting up more launch pads cost a lot. Yes. Will design and build on Raptor cost a lot (Musk said a $Bn when Congress asked, but I suspect the costs will change when SX has to fund it themselves). yes.

Quote
Start with assumptions about manufacturing costs, capex, capacity, material cost, variable costs per launch and so on and build on that.
What exactly are your goals here? The model simply says "with these assumptions this is the result."
If you don't like the numbers change them yourself.

If you think the model structure is wrong what would you rather it be?
Quote
Your model is something you might be able to use to deduct SpaceX's cost structures if you _knew_ all these other figures (cost per launch, overall profit, both of the for a certain number of flights per year, utilization, investment,...) which would then allow you to build on that but since you know none of them except for some vaguely defined gross figures you can't use it to deduct anything.
It's good enough to show (roughly) how prices could change with reuse and how (roughly) costs change with how many people you put on the refurbishment and wheather it flies enough to charge by the hrs actually needed or how many flights the LV does a year.

Some of those numbers are very "stiff." It takes a lot of flights before working out how many man hours are spent on a refurb becomes a viable way to charge for staff usage.

I'm interested in what the customer sees. Shotwell said at the Nato conference the full cost of a comm sat on SX is about $100m, same as an Ariane 5 ride share, but without having to find a 2nd comm sat of the right mass in close angular position in GEO to partner with.

IOW the cost reduction of flying SX for comm sats is zero.  :( It's more convenient but it's not $1 cheaper.

I'm looking at what can be deduced from public statements and their impacts on costs and prices charged.

What were you hoping it would tell you?

You've spent a long time explaining the shortcomings of the model. You clearly have a much better accountancy background than I do. The model is deliberately simplistic. A sanity check on price expectations, not to do SX's tax and benefits planning.  :)

Let me suggest you provide a better structure for answering the question in the thread title.

BTW I do understand things like gross and net profit and the idea of step changes in production costs bought on by exceeding capacity on existing plant (like say the diameter limit on say an FSW tank making machine) and break even to a point where "contribution" turns into pure profit (the attraction for newspaper and magazine publishers). IIRC for the original 707 it was around 60 aircraft sold. And yes I can run an IRR calculation if necessary.

Strictly speaking I'm not sure any LV has actually broken even without the national government of whatever state simply writing off the development costs (and often paying for the facilities as well it seems  :( ).

IOW Exactly what the UK and France did with Concorde and handing them to the (then) state owned airlines of each country.
« Last Edit: 07/07/2015 02:21 am by john smith 19 »
MCT ITS BFR SS. The worlds first Methane fueled FFSC engined CFRP SS structure A380 sized aerospaceplane tail sitter capable of Earth & Mars atmospheric flight.First flight to Mars by end of 2022 TBC. T&C apply. Trust nothing. Run your own #s "Extraordinary claims require extraordinary proof" R. Simberg."Competitve" means cheaper ¬cheap SCramjet proposed 1956. First +ve thrust 2004. US R&D spend to date > $10Bn. #deployed designs. Zero.

Offline pippin

  • Regular
  • Senior Member
  • *****
  • Posts: 2575
  • Liked: 312
  • Likes Given: 45
What exactly are your goals here? The model simply says "with these assumptions this is the result."
I don't think you understand the assumptions behind your model.
You say your model was about what reuse does to price and profit but in fact it does nothing like that because it does in no way take SpaceX' cost structure into account. The results of your model are just random figures due to that.

Quote
If you don't like the numbers change them yourself.
It's not about the numbers, it's the model that's wrong. It simply doesn't do what you seem to think it does.

Quote
If you think the model structure is wrong what would you rather it be?
C'mon, I wrote a _really_ long post about that.

Quote
It's good enough to show (roughly) how prices could change with reuse
No, it's not. That's what I've been trying to explain to you.
I took the effort to explain it, please take the effort to try to understand what I wrote.

Quote
Let me suggest you provide a better structure for answering the question in the thread title.
I can't because I don't have the necessary data. But there's no value in creating random figures based on totally wrong calculation models instead and call them calculations. Could just as well roll some dice.
Price and "profit" are irrelevant input values. You need to understand costs and costs structures to know whether reuse will actually reduce or increase said costs.
The comes understanding how reuse affects value for the customer (lower due to higher risk, lower flexibility? Higher due to lower risk? Same because it's just an insured flight? We don't know that, yet, neither does SpaceX or the customers) which in turn affects prices.
When you know that you'll know whether profits increase or not.
« Last Edit: 07/07/2015 04:15 am by pippin »

Offline ClaytonBirchenough

  • Full Member
  • ****
  • Posts: 777
  • ~ 1 AU
  • Liked: 34
  • Likes Given: 348
I'm not sure, but do investors have any impact on how quickly costs get amortized? I guess it really depends on what kind of investment it is, but from my understanding, in the case of SpaceX, Elon has controlling interest so it doesn't matter what any of the investors actually say (correct?)? Also, not sure if dividends are paid to investors (again speaking about SpaceX). I could see investors wanting to just buy stock to see the value of the company rise and then sell their stake for a profit. Not sure how all that ties in, but it might haha.

Also, do we know how much money Elon has personally put into developing F9R? Is it possible Elon personally paid for a large portion of F9R development so little/very little costs need to amortized over each flight?
Clayton Birchenough

Offline pippin

  • Regular
  • Senior Member
  • *****
  • Posts: 2575
  • Liked: 312
  • Likes Given: 45
Controlling stake will not help Elon because investors will demand their own terms for investing, something that's totally common for privately funded companies. They will be much better protected than Musk.

Dividends don't make sense in such a setup. Investors are not in for a cash flow but for a stake in the company and dividends would just reduce the cash available for the company they own anyway. You usually only start paying dividends when you generate so much cash that you can't use all of it for your growth anymore, SpaceX is far from there.

But SpaceX, as I understand, has also taken up debt which will have to pay interest.

All of this (except interest) isn't immediately important in the cost structure because investors only get paid in case of success, important aspects are ongoing investments they have to do (which require additional capital), personnel, machinery, real estate etc. a lot of which will be leased but nonetheless fixed. And then on the variable side contractors, materials and commodities sourced from outside suppliers. Non-commodities sourced from elsewhere fall somewhere in between, depending on the contract because you usually can't just buy more on a short-term basis and you usually have to guarantee certain quantities.

Offline ClaytonBirchenough

  • Full Member
  • ****
  • Posts: 777
  • ~ 1 AU
  • Liked: 34
  • Likes Given: 348
Controlling stake will not help Elon because investors will demand their own terms for investing, something that's totally common for privately funded companies. They will be much better protected than Musk.

Dividends don't make sense in such a setup. Investors are not in for a cash flow but for a stake in the company and dividends would just reduce the cash available for the company they own anyway. You usually only start paying dividends when you generate so much cash that you can't use all of it for your growth anymore, SpaceX is far from there.

But SpaceX, as I understand, has also taken up debt which will have to pay interest.

All of this (except interest) isn't immediately important in the cost structure because investors only get paid in case of success, important aspects are ongoing investments they have to do (which require additional capital), personnel, machinery, real estate etc. a lot of which will be leased but nonetheless fixed. And then on the variable side contractors, materials and commodities sourced from outside suppliers. Non-commodities sourced from elsewhere fall somewhere in between, depending on the contract because you usually can't just buy more on a short-term basis and you usually have to guarantee certain quantities.

So would a typical SpaceX investor today (speculation) be getting paid by a fix rate determined initially upon investing in SpaceX? Or is it through the rise in the price of the stock they hope to get returns on?
Clayton Birchenough

Offline john smith 19

  • Senior Member
  • *****
  • Posts: 10351
  • Everyplaceelse
  • Liked: 2431
  • Likes Given: 13606
I don't think you understand the assumptions behind your model.
You say your model was about what reuse does to price and profit but in fact it does nothing like that because it does in no way take SpaceX' cost structure into account. The results of your model are just random figures due to that.
Since you seem to know my spreadsheet better than I do why don't you tell me what the assumptions are?
Quote
C'mon, I wrote a _really_ long post about that.
Reading it was rather like one of those IQ test questions with 2 paragraphs of dense prose and a last line that says "The above is irrelevant."  :(

In your case telling me what I hadn't done and then putting essentially "But I don't know those numbers either." We are both ignorant of the numbers. I guess that puts us in exactly the same position as >99% of the readers here.  :)

Again this sheet is all around the F9 and it's possible 1st stage reuse. You seem to want it to tell you wheather it will generate enough money to keep SX alive. I'm not interested in that question. I'm interested in what reuse can do to the price to the customer.

Quote
I can't because I don't have the necessary data. But there's no value in creating random figures based on totally wrong calculation models instead and call them calculations. Could just as well roll some dice.
Price and "profit" are irrelevant input values. You need to understand costs and costs structures to know whether reuse will actually reduce or increase said costs.
I find customers have quite a strong interest in price when there's little to differentiate the suppliers. When they're all good (or all equally expensive) not so much.  :(

I find companies usually develop quite an interest when the price they charge and the volumes they sell don't cover all their costs.
Quote
The comes understanding how reuse affects value for the customer (lower due to higher risk, lower flexibility? Higher due to lower risk? Same because it's just an insured flight? We don't know that, yet, neither does SpaceX or the customers) which in turn affects prices.
When you know that you'll know whether profits increase or not.
I clearly seem to have offended you by calling this a model. You also seem to want it to answer much wider questions about SX's financial health. It's focus is solely on the F9SR.

How about I call it a game instead ? It's simply trying to collect together the public statements SX staff have made and see where they go.

You seem to have expected a first principles cost analysis of what SX have spent, what they are spending and what they are committed to spending.

My apologies. This is not that kind of spreadsheet.

The game tries to identify as many known cost items. Not their value, just there existence. In some cases it uses public figures to set initial values for those cells.

I'd welcome anyone with better numbers for range costs for example. Likewise do people want a "first principals" propellant cost estimate based on tank volume or will they settle for Musks number of "about $120k" for an F9 flight?

The game allows different "players" (anyone whose downloaded it) to decide what they think will be needed to refurbish a stage and how many times it can be done to see what the outcomes are if they vary those assumptions?

Note the game makes minimal assumptions about reuse. It assumes that refurb will take a)Staff and b)Stuff (the "arbitary refurb costs" item) to produce a resale price for the LV. There will a "range cost" to support range hardware and range dedicated staff.

At some level of staff and stuff use and reflights it does become cheaper to just build a new one, which I think is a pretty good model of real life.   :(

The game makes no comment on wheather what's left over from the expected costs (and their sizes) is enough to support SX, or even part of it. It makes no comment what is the break even quantity of launches SX requires to cover all that new investment, either back to banks or equity investors.

The games purpose is to let players see how "stiff" some of those numbers are to change, giving people a feel about what it's worthwhile trying to improve, which in turn will depend on what peoples starting assumptions are.
« Last Edit: 07/07/2015 12:46 pm by john smith 19 »
MCT ITS BFR SS. The worlds first Methane fueled FFSC engined CFRP SS structure A380 sized aerospaceplane tail sitter capable of Earth & Mars atmospheric flight.First flight to Mars by end of 2022 TBC. T&C apply. Trust nothing. Run your own #s "Extraordinary claims require extraordinary proof" R. Simberg."Competitve" means cheaper ¬cheap SCramjet proposed 1956. First +ve thrust 2004. US R&D spend to date > $10Bn. #deployed designs. Zero.

Offline pippin

  • Regular
  • Senior Member
  • *****
  • Posts: 2575
  • Liked: 312
  • Likes Given: 45
Since you seem to know my spreadsheet better than I do why don't you tell me what the assumptions are?
I did
Quote from: pippin
your cost model assumes a fixed cost structure with only variable costs and no changes to cost structure with utilization
What that means is: you assume that if SpaceX flies, say, 15 times and now has to build one less 1st stage because they can reuse it, they will save 1/15th (times cost share of Stage 1) of the production cost.
But that's not the case. Depending on their contracts and their cost structure they will likely only save a small part of that (something like 25%-30% would be my guess, but it's really just that, a guess), they might not save anything at all.

Now, if they don't save a lot that obviously has a lot of impact on the resulting business model for reuse since bringing back the core will create some costs of it's own.

That means that flight rates do already have an impact on the business model of the very first stage they are reflying.

If they change the metrics by more than 1/15th of the production run this obviously might have a much bigger impact. Say their factory is scaled to produce 20 cores a year and now due to reuse they only do 10, then these 10 will likely be much more expensive per core than the 20 the factory was scaled for. They might have to lay off people, re-negotiate supplier contracts and so on. This is what I meant when I said that a fixed cost structure is unrealistic.

You also talk about "profits" (it's in your model), these profits will also depend on the customer side of things (how the customer perceives the reuse) and this probably also will be affected by flight rates, market conditions etc. which makes it even less predictable which is why I said I'd focus on cost alone.

Quote
Again this sheet is all around the F9 and it's possible 1st stage reuse. You seem to want it to tell you wheather it will generate enough money to keep SX alive.
Not at all. I had an even more focussed question than you because I only looked at the _cost_ effect of said 1st stage reuse, dropping that "profit" part which is inevitably tied to the rest of what SpaceX does.
You were bringing in the rest of their operations by adding that "profit" figure in there.
And before you ask: I did explain why that is in my last post already...
« Last Edit: 07/07/2015 01:49 pm by pippin »

Offline MikeAtkinson

  • Full Member
  • ****
  • Posts: 1980
  • Bracknell, England
  • Liked: 784
  • Likes Given: 120
Since you seem to know my spreadsheet better than I do why don't you tell me what the assumptions are?
I did
Quote from: pippin
your cost model assumes a fixed cost structure with only variable costs and no changes to cost structure with utilization
What that means is: you assume that if SpaceX flies, say, 15 times and now has to build one less 1st stage because they can reuse it, they will save 1/15th (times cost share of Stage 1) of the production cost.
But that's not the case. Depending on their contracts and their cost structure they will likely only save a small part of that (something like 25%-30% would be my guess, but it's really just that, a guess), they might not save anything at all.

We know there is a big discontinuity in SpaceX's costs at around 40 cores per year. Above that Hawthorne is not large enough for core production. Setting up a second manufacturing site with room to grow is going to be very expensive. This alone makes the 'model' invalid.

A second point is that no model no matter how good will give useful results when fed wrong input data. We have no insight into SpaceX costs, encouraging people to put their own numbers in only produces more wrong results.

A related point is that models need to be validated, we have no cost/price data to validate the model.

Offline john smith 19

  • Senior Member
  • *****
  • Posts: 10351
  • Everyplaceelse
  • Liked: 2431
  • Likes Given: 13606
We know there is a big discontinuity in SpaceX's costs at around 40 cores per year.
I did not know that. Not surprising, given we saw that 8 cores a year as about the right level for ULA's Decataur facility.
Quote
Above that Hawthorne is not large enough for core production. Setting up a second manufacturing site with room to grow is going to be very expensive. This alone makes the 'model' invalid.
No, it tells people either SX had better get their reuse plan sorted out before they try to launch 40cores in a year or they are going to need a shedload of money to set up a new factory.
Quote
A second point is that no model no matter how good will give useful results when fed wrong input data. We have no insight into SpaceX costs, encouraging people to put their own numbers in only produces more wrong results.
It's already shown me that it's very doubtful that SX will get savings some people are expecting.
MCT ITS BFR SS. The worlds first Methane fueled FFSC engined CFRP SS structure A380 sized aerospaceplane tail sitter capable of Earth & Mars atmospheric flight.First flight to Mars by end of 2022 TBC. T&C apply. Trust nothing. Run your own #s "Extraordinary claims require extraordinary proof" R. Simberg."Competitve" means cheaper ¬cheap SCramjet proposed 1956. First +ve thrust 2004. US R&D spend to date > $10Bn. #deployed designs. Zero.

Offline Jim

  • Night Gator
  • Senior Member
  • *****
  • Posts: 37442
  • Cape Canaveral Spaceport
  • Liked: 21452
  • Likes Given: 428

Not surprising, given we saw that 8 cores a year as about the right level for ULA's Decataur facility.


Not true.  ULA launch rate is constrained by pads and not core production.

Offline ClaytonBirchenough

  • Full Member
  • ****
  • Posts: 777
  • ~ 1 AU
  • Liked: 34
  • Likes Given: 348
Not true.  ULA launch rate is constrained by pads and not core production.

How is it constrained by pads?
Clayton Birchenough

Offline spacenut

  • Senior Member
  • *****
  • Posts: 5181
  • East Alabama
  • Liked: 2587
  • Likes Given: 2895
The Decautar facility was originally designed for 40 cores a year. 

Offline Jim

  • Night Gator
  • Senior Member
  • *****
  • Posts: 37442
  • Cape Canaveral Spaceport
  • Liked: 21452
  • Likes Given: 428

How is it constrained by pads?

Atlas V is constrained by VIF/MLP turn around. 

Offline ClaytonBirchenough

  • Full Member
  • ****
  • Posts: 777
  • ~ 1 AU
  • Liked: 34
  • Likes Given: 348

How is it constrained by pads?

Atlas V is constrained by VIF/MLP turn around.

Thanks.

So is 8 flights a year currently the maximum flight rate Atlas V is constrained to due to pads with existing workforce/facilities? Or 8 flights a year is not the maximum, and you were just pointing out the limiting factor on Atlas V launches would be pad turn around not core production at the Decautar facility?
Clayton Birchenough

Offline Jim

  • Night Gator
  • Senior Member
  • *****
  • Posts: 37442
  • Cape Canaveral Spaceport
  • Liked: 21452
  • Likes Given: 428
Or 8 flights a year is not the maximum, and you were just pointing out the limiting factor on Atlas V launches would be pad turn around not core production at the Decautar facility?

correct

Offline JamesH

  • Full Member
  • ****
  • Posts: 525
  • United Kingdom
  • Liked: 284
  • Likes Given: 7
We know there is a big discontinuity in SpaceX's costs at around 40 cores per year.
I did not know that. Not surprising, given we saw that 8 cores a year as about the right level for ULA's Decataur facility.
Quote
Above that Hawthorne is not large enough for core production. Setting up a second manufacturing site with room to grow is going to be very expensive. This alone makes the 'model' invalid.
No, it tells people either SX had better get their reuse plan sorted out before they try to launch 40cores in a year or they are going to need a shedload of money to set up a new factory.
Quote
A second point is that no model no matter how good will give useful results when fed wrong input data. We have no insight into SpaceX costs, encouraging people to put their own numbers in only produces more wrong results.
It's already shown me that it's very doubtful that SX will get savings some people are expecting.

From SpaceX's POV, what other people think is irrelevant. It's what their plans and spreadsheets say that is important. Musk and his crew are not stupid people.

Offline pippin

  • Regular
  • Senior Member
  • *****
  • Posts: 2575
  • Liked: 312
  • Likes Given: 45
No, it tells people either SX had better get their reuse plan sorted out before they try to launch 40cores in a year or they are going to need a shedload of money to set up a new factory.
Exactly. Which is pretty much equivalent to "depending on flight rate" :)

Quote
It's already shown me that it's very doubtful that SX will get savings some people are expecting.
Well, I believe _that_ is true, too.
I don't have a good model but given how much fixed costs they have to have with their high vertical integration it's highly unlikely that they save more than 10-20% per flight and it's definitely not in the range of "orders of magnitude cheaper".

Except of course in the very case you just mentioned. With very high flight rates that would require additional capacity opportunity costs would go up quite a bit (but building 80 cores a year in two factories would of course _still_ be cheaper per core than building 40 in one).

So it's back to square one: like for the Shuttle, everything depends on the flight rate...

Offline spacenut

  • Senior Member
  • *****
  • Posts: 5181
  • East Alabama
  • Liked: 2587
  • Likes Given: 2895
Musk said 50% cost reduction if reusing a core.  Say it is only 75% after refurbishment and a new second stage.  If the core can be used 4-5 times, then it would average about 50%.  Until he gets a core back to see what stresses are put on it we will not know.  If he can reuse it 4 times, the 5th time could be expendable.  This would give him 160 launches per year.  That is 3 per week.  With the Kennedy facilities, along with Texas and California, I think it could be done.  He could charge a reduced cost per time it was used.  If he gets 10 satellites per launch with the WiFi grid, that is 1600 satellites.  This being ideal conditions.  It is possible, but figure about half to be realistic, also maybe only one to two reuses. 

Offline john smith 19

  • Senior Member
  • *****
  • Posts: 10351
  • Everyplaceelse
  • Liked: 2431
  • Likes Given: 13606

Not surprising, given we saw that 8 cores a year as about the right level for ULA's Decataur facility.


Not true.  ULA launch rate is constrained by pads and not core production.
Did not write "constrained," wrote "right."


The Decautar facility was originally designed for 40 cores a year.
Yes, but my figure for 8 was what ULA told the USAF they'd like per year for their block buy.

In production engineering terms that would be the "Economic Batch Quantity" for the facility.

At that level all staff are employed in normal hours and all holidays can be taken, no overtime has to be paid and there is plenty of slack for late deliveries to be accommodated from internal stores. IOW minimal costs spread across maximum product.

But outside that your starting to work multiple shifts, paying overtime and cancelling holidays to use your core staff. Then you're bringing back retirees and people you've laid off as they already know their jobs. By now I'd guess your up to 2x to 3x you're comfortable production level with a regular shift.

Then you're looking at bringing in agency staff to cover to the 4x and 5x level. At this point you're running flat out. In extreme cases you'll be spinning up some in house training programmes to teach them to do certain jobs or upskilling and promoting some of your permanent staff.
And since this is the rocket business and National Security programmes a proportion of those will need to be security vetted to the relevant government standards before training.

I'm sure Decautar can do 40 cores a year.

But ramping up to do that level will not be cheap.

Well, I believe _that_ is true, too.
I don't have a good model but given how much fixed costs they have to have with their high vertical integration it's highly unlikely that they save more than 10-20% per flight and it's definitely not in the range of "orders of magnitude cheaper".
Actually I think this is where SX have turned conventional industry thinking on its head.

Historically bigger players have tried to move to a "minimal core skills" model, buying in everything else.

But launch volume contraction, M&A in the supply chain and the very high cost of getting a new part "space certified" means that (now) you might have a single mfg that makes a relevant component. They make a small number of them a year. Their prices are exhorbitant by more conventional industries standards and even then there QC may not be that great as the inspectors only have to check a few of these a year as well.

And should you require the slightest change in spec they'll want even more cash. :(
(What the CEO of AmRoc called "charging for change.")

But once you've got a set of space simulation and test facilities in house provided they're big enough you can do your own "space certification" testing (at least for your own LV). Then the cost per new pat becomes incremental rather than a step change for every component.
Quote
So it's back to square one: like for the Shuttle, everything depends on the flight rate...
And how much of each launch you throw away.

The calculation is not that hard to do. This has been known in industry for a very long time.

That's why the real step change is a vehicle that delivers the same payload as an equal sized ELV.
MCT ITS BFR SS. The worlds first Methane fueled FFSC engined CFRP SS structure A380 sized aerospaceplane tail sitter capable of Earth & Mars atmospheric flight.First flight to Mars by end of 2022 TBC. T&C apply. Trust nothing. Run your own #s "Extraordinary claims require extraordinary proof" R. Simberg."Competitve" means cheaper ¬cheap SCramjet proposed 1956. First +ve thrust 2004. US R&D spend to date > $10Bn. #deployed designs. Zero.

Offline joek

  • Senior Member
  • *****
  • Posts: 4870
  • Liked: 2783
  • Likes Given: 1097
Yes, but my figure for 8 was what ULA told the USAF they'd like per year for their block buy.

Nit: Original ULA proposal was based on 10 cores/yr (50 total); USAF committed to 36 from ULA with 14 open.

Tags:
 

Advertisement NovaTech
Advertisement Northrop Grumman
Advertisement
Advertisement Margaritaville Beach Resort South Padre Island
Advertisement Brady Kenniston
Advertisement NextSpaceflight
Advertisement Nathan Barker Photography
1