Incentive for Space Stations

[linuxgurugamer]

1 minute ago, tater said:

Yours is exactly the same as Giscard's. Servicing an unmanned satellite. What the sat does doesn't matter. The only real world example had repair cost more than replacement.

Note that a station is likely in LEO because of radiation hazards, and solar power is likely in GEO. So such a maintenance requires an equatorial station ideally, and then require sending workers TO the power sat. Why would you not just send them there directly, instead of keeping them on a station, then sending a spacecraft that will certainly require fuel deliveries anyway.

Yes, but a Rectanna will be huge (something which is measured using miles or kilometers, not meters) and cost billions of dollars.  Rockets take time to prep and launch, for routine maintenance, having an in-orbit station would be extremely useful.

And, you are now quibbling.  You asked for examples.  They have been given to you.  The fact that YOU don't think they make sense doesn't make them any less valid, and frankly, unless you are actually in the industry, you are basically spouting nonsense when you start talking economics of things like this.

[Jovus]

@linuxgurugamer You should check out the Manned Orbiting Laboratory program and why it was cancelled.

 

Long story short, for large, complex, vital-to-function satellites it's still cheaper to send up an engineer or a spare satellite if there are problems than to keep an engineer on orbit just in case.

[tater]

3 minutes ago, linuxgurugamer said:

Yes, but a Rectanna will be huge (something which is measured using miles or kilometers, not meters) and cost billions of dollars.  Rockets take time to prep and launch, for routine maintenance, having an in-orbit station would be extremely useful.

And, you are now quibbling.  You asked for examples.  They have been given to you.  The fact that YOU don't think they make sense doesn't make them any less valid, and frankly, unless you are actually in the industry, you are basically spouting nonsense when you start talking economics of things like this.

I'm not quibbling, it's exactly the same, satellite service.

 

How is it economic to keep people in LEO 24/7/365 so that you can occasionally send them up on short duration missions to GEO?

[TheEpicSquared]

@Jovus Actually, your idea about quarantine is a really good one. And safe, too - certainly safer than doing it on earth.

[tater]

From a KSP standpoint, construction of something huge doesn't require a station, they can just live in a pod while they work, if you even need kerbals. If kerbals had AI, and could be tasked to do things, then orbital construction might be a thing. You'd send kerbals to supply a depot with parts, then kerbals based at the depot would build whatever they were taken to build. As it is, you really need only 1 with KIS, or a probe core otherwise to dock stuff together.

Seems like the major costs would be transportation, not equipment, so if you are sending something to GEO, you might as well bring the spare as SOP. The only reason to bring the broken bit down would really be to deorbit it. I'd think that telepresence would be superior to risking people in a high radiation environment, as well.

Edited September 13, 2016 by tater

[wrobel-cwirek]

11 hours ago, Jovus said:

More serious suggestion: quarantine.

Some disease/device/phenomenon needs study, but it's just too dangerous to study it anywhere on Earth, despite strict quarantine procedures, because the potential downside of a quarantine break is too high. You locate the lab on a space station, and if a containment breach happens the whatever-it-is has to both survive re-entry and have the delta-V to pull it off.

Problem is with sending the bacteria/wirus samples to the station - in case of ascent failure there is no guarantee of destruction of the hazardous cargo (as in the Columbia shuttle catastrophe were some worms from bio-samples survived the crash). With viruses/bacteria it would be even more probable.

And then you have Day-Z

 

Edited September 14, 2016 by wrobel-cwirek

[Codraroll]

The crux of the matter is still KSP. The game. It's currently feasible to dock parts together and make huge space stations. There are plenty of parts designed for doing exactly that. But apart from looking really cool, and providing lots of repeated docking training, you have little reason to build space stations. Even the Science Lab can be trivially launched as a single vessel and work perfectly; you don't even need to put a docking port on it.

Comparisons to the real world could be helpful to the discussion, but not if we forget to keep it in the persective of KSP. Economic feasibility is hardly a hindrance in KSP, you're showered with money you don't need anyway, as rewards from other simple missions. Instead of asking "Why would anybody in the real world pay for a space station?", we should ask "Why would anybody in the real world want a space station?" Screw the economics, those aren't problems in KSP unless you specifically make it so. It's a game, and it can afford to boldly jump over most hindrances that would make solid arguments against real-life space stations.

So, rattling off a few things I think it would be really cool to have in space:

• Construction yards for vacuum-only ships.
• Test facilities for engines, spacesuits, tools, hull structures or other large-scale components in zero gravity and hard vacuum. A real-world space economy would have plenty of companies making products for space usage, and the need for a certification body such as TÜV would arise. I've worked briefly in a place certifying products for the building industry, and they had plenty of work to do all year 'round.
• Zero-G movie studios.
• Zero-G sports complexes (imagine sports like handball or gymnastics without gravity).
• Orbital greenhouses - sunlight 'round the clock and no gravity keeping plants down. It'd be expensive food, but hey, Kerbin probably has gourmet restaurants that would cough it up.
• Orbital hotels, spas and conference centers (hey, might as well go the extra mile).
• Orbital science labs or even education facilities - face it, you'd have loved to do a few weeks as an exchange student in low Earth orbit.
• Biological and zoological test facilities. How do birds handle zero-G? Dogs? Cats? Fish? How do they handle vacuum?
• Test facilities for artificial gravity solutions.
• A "bus stop" atop the gravity well. Ships designed for asteroid mining, deep space exploration or interplanetary shuttles wouldn't like diving into Kerbin's atmosphere, nor would they ever get out if they tried. A shuttle would be required to cycle crews on those ships. Rather than every space company launching its own crews to ships in orbit, they could dock their ships at the space station and use the station's shuttle service to transport crew and goods between orbit and the surface.

Most of these are feasible given Kerbal economics, and could provide a certain amount of fun gameplay. Don't let the conventions of real-world affordability get in the way of fun. Ask why you want a space station before considering whether you'd pay for it.

[wrobel-cwirek]

59 minutes ago, Codraroll said:

Zero-G sports complexes (imagine sports like handball or gymnastics without gravity).

Ender's Game

Edit:

To add to the discussion - we need whole game quest system redone (less of the "test X part at Y speed at Z height"), and more the proper quests like sending probes to other planets, creating (and then maintaining, and also cancelling space stations, building new one, changing crews, etc. - liek the mode for Stations and Bases), sending crews to different planets etc., building complex vacuum-only ships on orbit and reusing them - like quest line to build ship, send it crewd to Duna, go back, refuel, change crew, send it to Eve, etc.  With proper balanced rewards.

I would love... love to have quests in science mode too... just without $$$ rewards.

Also - whats it the point of prestige? I don't see any for it. Also administration building sux... big time. This should support better the game somehow. Like the Strategies mode do. The stock "strategies" are very hard to understand. I just don't care about them, The Strategies mode is better. 

But for this we need at least proper TWR/dV display in stock game, and stock transfer window planner. This two are a must - so planning you can do on paper.

Maybe also mission planner (like you can have white-board where you can put steps you want to do a mission (like start from Kerbin, transfer to Duna, Duna landing, Duna start, transfer back to Kerbin, and it would help show what are the time for this, the dV, TWR needed, etc. etc. Maybe with "sub-missions" if for example you want to build mother ship with "drones" and "landers".

Edited September 14, 2016 by wrobel-cwirek

[tater]

3 hours ago, Codraroll said:

The crux of the matter is still KSP. The game. It's currently feasible to dock parts together and make huge space stations. There are plenty of parts designed for doing exactly that. But apart from looking really cool, and providing lots of repeated docking training, you have little reason to build space stations. Even the Science Lab can be trivially launched as a single vessel and work perfectly; you don't even need to put a docking port on it.

Comparisons to the real world could be helpful to the discussion, but not if we forget to keep it in the persective of KSP. Economic feasibility is hardly a hindrance in KSP, you're showered with money you don't need anyway, as rewards from other simple missions. Instead of asking "Why would anybody in the real world pay for a space station?", we should ask "Why would anybody in the real world want a space station?" Screw the economics, those aren't problems in KSP unless you specifically make it so. It's a game, and it can afford to boldly jump over most hindrances that would make solid arguments against real-life space stations.

Yeah, I agree, my off topic tangent was because I was saying that the lack of economic drivers in game wasn;t unrealistic, as there really are no economic drivers, stations are for other reasons.

 

3 hours ago, Codraroll said:

So, rattling off a few things I think it would be really cool to have in space:

• Construction yards for vacuum-only ships.
• Test facilities for engines, spacesuits, tools, hull structures or other large-scale components in zero gravity and hard vacuum. A real-world space economy would have plenty of companies making products for space usage, and the need for a certification body such as TÜV would arise. I've worked briefly in a place certifying products for the building industry, and they had plenty of work to do all year 'round.

Yeah, these two are great, although they come with a standard KSP problem---Kerbals are useless by themselves, so the player has to hand-deliver all the supplies needed. I'm all-in for construction yards, though. 

 

3 hours ago, Codraroll said:

• Zero-G movie studios.
• Zero-G sports complexes (imagine sports like handball or gymnastics without gravity).

Fun, but I'm unsure how this works in game. Honestly, the "tourism" could take care of these with better contracts (and less random).

 

3 hours ago, Codraroll said:

• Orbital greenhouses - sunlight 'round the clock and no gravity keeping plants down. It'd be expensive food, but hey, Kerbin probably has gourmet restaurants that would cough it up.

Glass in large quantities really isn't a thing, growing stuff would likely be under electric light, powered by solar, or via a heliostat directing light into a more enclosed (and easier to regulate) greenhouse.

 

3 hours ago, Codraroll said:

• Orbital hotels, spas and conference centers (hey, might as well go the extra mile).

That one is out there, but yeah, the contract system need to have hotels be a thing.

3 hours ago, Codraroll said:

• Orbital science labs or even education facilities - face it, you'd have loved to do a few weeks as an exchange student in low Earth orbit.
• Biological and zoological test facilities. How do birds handle zero-G? Dogs? Cats? Fish? How do they handle vacuum?
• Test facilities for artificial gravity solutions.

Again, unsure of the mechanism for incentive in KSP. I have suggested that certain parts require specific "science," so this could be a thing... build a certain facility, and run it for XXX days to unlock new parts.

3 hours ago, Codraroll said:

• A "bus stop" atop the gravity well. Ships designed for asteroid mining, deep space exploration or interplanetary shuttles wouldn't like diving into Kerbin's atmosphere, nor would they ever get out if they tried. A shuttle would be required to cycle crews on those ships. Rather than every space company launching its own crews to ships in orbit, they could dock their ships at the space station and use the station's shuttle service to transport crew and goods between orbit and the surface.

Most of these are feasible given Kerbal economics, and could provide a certain amount of fun gameplay. Don't let the conventions of real-world affordability get in the way of fun. Ask why you want a space station before considering whether you'd pay for it.

This is just logistical, and I think many of us do this anyway---if for no other reason than we want to use our stations, for something.

[Jovus]

5 hours ago, wrobel-cwirek said:

Problem is with sending the bacteria/wirus samples to the station - in case of ascent failure there is no guarantee of destruction of the hazardous cargo (as in the Columbia shuttle catastrophe were some worms from bio-samples survived the crash). With viruses/bacteria it would be even more probable.

And then you have Day-Z

 

Indeed. There are a few ways this might not be a problem, though:

• The ingredients are inert until combined
• Atmospheric transmission isn't a vector
• The thing came from space in the first place

In the second case, you still get the problem of your MacGuffin surviving the crash and being introduced to the local environment for X minutes/hours until the cleanup crew can contain it - if they even can.

[Aegolius13]

I like the idea of a "hotel" part - a crew module that's design for long term habitation, with contracts to put either astronauts or Kerbals there for a certain length of time.  And make it bulky enough, heavy enough, and/or demanding enough on electricity, that it would take a true space station or base to support it.  This would kind of emulate the higher needs for life support without going into all the detail that the mods do. 

Also, I think they still do not have the mechanics of the Mobile Processing Lab quite right, but I'm not sure exactly what the fix is. If I were starting from scratch, I might allow it to transmit samples back to Kerbin without penalty, but without the crazy multiplier it has now (two guys in a module can learn more about a sample than all the scientists back home? ).  And then maybe have specific contracts to bring back certain samples to the MPL, with a reasonable science reward.  

[tater]

Honestly, we don't need a "unitasker" part. Add a new, larger habitat part or parts, and deal with the "hotel" issue in contracts.

Stations can already get new contracts to add to them. Add a flag such that if a station is built as part of a "hotel" contract, then it receives guest contract. Done.

[Workable Goblin]

On September 13, 2016 at 3:05 PM, tater said:

I'm not quibbling, it's exactly the same, satellite service.

It really isn't. Scale is not just a minor detail here, it's completely dominant. A good comparison is computers on Earth. Suppose you have a cheap smartphone, something that costs $100 or less. If it breaks it probably makes more sense for you to get a new one than to try to repair it, just because it would cost nearly as much to fix as it did for you to buy it in the first place. The same is true of cheap tablets, cheap laptops, and other cheap consumer computers. Now compare to someone who has a big, dedicated cluster or mainframe or similarly high-powered computers. In that case, fixing the computer if it breaks totally makes sense; that kind of hardware costs millions of dollars and months or years to put together, and often does millions of dollars worth of "work" every day (whether that be scientific computations or processing financial transactions or any other computationally demanding task), whereas it costs much less and takes much less time to repair it. Many such computers even have dedicated staff on hand who can fix problems as they arise, which would be obviously silly for even very expensive consumer gear. Obviously different conditions apply, such that saying "it's exactly the same, computer service!" is completely wrong and misleading.

More or less is true of satellites today versus the kind of giant solar power satellite that linuxgurugamer is talking about. Satellites today aren't cheap, exactly, but they have a construction and launch cost which is comparable to the cost of servicing, and it would probably take about as long to put together a servicing mission as to build and launch a new satellite, nowadays. So it doesn't make any economical sense to service satellites. A huge solar power satellite, though, would be enormously costly and require a long time to build, not to mention that there's physically less room available for such satellites than for modern communications satellites (and since they want to live at GEO, space is actually a concern). When combined with the fact that building such satellites economically would necessarily require a large, costly infrastructure that would make it cheaper to travel to GEO and LEO in the first place, it it much more likely that it would make sense to repair such a satellite than to repair modern-day satellites, just as it makes more sense for big companies to repair the giant computers they use for processing-intensive tasks than for consumers to repair their tablets and smartphones and laptops and so on, perhaps even to the point where it makes sense to permanently station maintenance staff at the satellite in case of any problems.

[tater]

@Workable Goblin We only have a single example of repair, and while HST was hugely expensive, it still would have been cheaper to throw it away and send a new one. Even with the repair, that repair was a mission by a spacecraft, it was not some sort of extemporaneous repair from a station. Any such repair requires that all spares are on hand already. If not, it's a dedicated mission, and doesn't require the station.

It is functionally the same, as for a space station, you need to demonstrate that it's economical to have all possible spares on hand, as well as crew, on the off chance you need a repair. Then you need to spend a huge amount of dv to get the repair crew and parts there. If you have a craft at station for this, you likely have propellant boil off, so constant propellant resupply as well.

Why would that be cheaper than sending the repair crew and specific parts needed from planetside?

 

Edited September 14, 2016 by tater

[Jonfliesgoats]

Think Hubble repairs v. repairs to the Lincoln Calibration Sphere (LCS has to have been hit by a micrometeorite by now).

[tater]

Clearly a power station with a huge cross sectional area will have more interactions with micrometeorites, etc than satellites do. The question is how often would it be worthwhile to go all the way to GEO to fix it. The repair craft would have to be fairly substantial, and be able to haul whatever repair materials are required, for as many days as such repairs might take. I could certainly see that craft as being useful to leave in LEO, and refuel, and resupply with repair parts. Why would it need a station to attach to? Every XX months, you fly a refuel/crew mission to the repair vessel, then it heads to GEO, does repairs, and returns to LEO. Perhaps it hauls along the crew delivery capsule, and the crew returns to Earth.

Does anyone really envision a station full of repair crew "on call" to pop up to GEO to fix a bad PV element, then they head back to LEO, then head up a couple days later when another goes out?

[Workable Goblin]

15 hours ago, tater said:

@Workable Goblin We only have a single example of repair, and while HST was hugely expensive, it still would have been cheaper to throw it away and send a new one. Even with the repair, that repair was a mission by a spacecraft, it was not some sort of extemporaneous repair from a station.

That's not actually true, in two ways. First and least interestingly, there were a number of satellite servicing missions before Hubble was even launched; SolarMax, Palapa B2, Westar 6, Syncom IV-3, and Intelsat 603 all had Shuttle missions partially or entirely dedicated to dealing with their malfunctions, all of which were at least partially successful. Of course, except for Intelsat 603 all of those were pre-Challenger missions, so they don't say much about the economics, but they happened.

Second, it just isn't true that it would have been cheaper to throw Hubble away and send a new one. The overall Hubble program, including servicing missions, has cost about $10 billion to date. $4 billion of that was for construction and launch, which means each service mission averaged around $1 billion in marginal costs (including instruments, launch, extra operations time, and so on). Major observatories like Hubble tend to take fifteen to twenty years to develop and launch, and I can't find any that cost less than $3 billion (Chandra). The one telescope that was actually developed as a Hubble replacement, James Webb, has had infamous budgetary overruns and is currently projected to cost around $9 billion for construction, launch, and five years of operations, not significantly less than Hubble itself with servicing. It wasn't possible to just build and launch a second Hubble in 1991, and anyway it wouldn't have been desirable, since astronomy and technology had moved on since the design was finalized in the 1970s and the astronomers would inevitably have wanted new instruments. Assuming that they would just build a second Hubble and send it up on the next shuttle (which a lot of people who say this seem to think) is extremely unrealistic; launching clones of particular vehicles has been very rare since the 1970s due to increasing costs and build times.

(And yes, all those are real dollars and hence comparable) 

15 hours ago, tater said:

It is functionally the same, as for a space station, you need to demonstrate that it's economical to have all possible spares on hand, as well as crew, on the off chance you need a repair. Then you need to spend a huge amount of dv to get the repair crew and parts there. If you have a craft at station for this, you likely have propellant boil off, so constant propellant resupply as well.

For the case under consideration (servicing GEO power satellites), all of that is quite reasonably economical. Such satellites are much too big to build on the ground and launch into space, so they're going to need assembly stations in orbit to put them together. Such stations will, by definition, have a lot of satellite parts on board--spares, in other words. Such satellites are in fact of marginal economic viability for any reasonable (and many unreasonable) reductions in launch costs from Earth, so they very nearly require space-based manufacturing to make sense. In that case, obviously the spare parts are made in space, so it would likely be more costly to fly them down to Earth to supply Earth-based servicing missions. Since such satellites require infrastructure that makes moving people and supplies through considerable delta-Vs cheap and easy, it wouldn't be much trouble to ship them around for repair missions, either. And as far as propellant boil-off...well, use storable propellants. Sure, they're toxic, but that just means you have to isolate them from crew quarters. Such services might also come along with other roles. For example, the spacecraft used to ship crews and parts up to satellites that need repairing might also launch satellites and beyond Earth orbit missions, or the station might also be a propellant depot that stores enough fuel and oxidizer to ship anything to anywhere anyway.

[tater]

49 minutes ago, Workable Goblin said:

Second, it just isn't true that it would have been cheaper to throw Hubble away and send a new one. The overall Hubble program, including servicing missions, has cost about $10 billion to date. $4 billion of that was for construction and launch, which means each service mission averaged around $1 billion in marginal costs (including instruments, launch, extra operations time, and so on). Major observatories like Hubble tend to take fifteen to twenty years to develop and launch, and I can't find any that cost less than $3 billion (Chandra). The one telescope that was actually developed as a Hubble replacement, James Webb, has had infamous budgetary overruns and is currently projected to cost around $9 billion for construction, launch, and five years of operations, not significantly less than Hubble itself with servicing. It wasn't possible to just build and launch a second Hubble in 1991, and anyway it wouldn't have been desirable, since astronomy and technology had moved on since the design was finalized in the 1970s and the astronomers would inevitably have wanted new instruments. Assuming that they would just build a second Hubble and send it up on the next shuttle (which a lot of people who say this seem to think) is extremely unrealistic; launching clones of particular vehicles has been very rare since the 1970s due to increasing costs and build times.

(And yes, all those are real dollars and hence comparable) 

Untrue. Marginal costs are meaningless. Total program cost, divided by launches. Every launch cost around 1.5 billion using that measurement in constant dollars. Had Shuttle not been a thing, HST goes via Titan III? likely (or another variant, this is counterfactual, so they'd make what they needed). The launch is less than 1/2 what the shuttle launch added to HST, and there are not X repair missions at a cost of 1.5B each. If you imagine a more limited lifespan, you take the opportunity cost of not blowing billions a year on Shuttle and have new (better) HSTs popping out every 10 years or so.

So HST could have launched on Titan, as other Keyholes did---HST and KH are un-coincidentally pretty much exactly the same (it's why the Shuttle cargo bay was the size it was, and why HST had the mirror it had instead of the planned 3m size). HST was to be launched in '86 (delayed from '85), had it been launched with Titan, they would have discovered the flaw in 1986, and would have had 7 years to build a new one and still come in at the same time HST became useful in 1993 with the COSTAR repair.

Bottom line is that it certainly could have been replaced a few times for the multi-billion dollar cost of what, 5-6 total shuttle missions for HST?

 

Quote

For the case under consideration (servicing GEO power satellites), all of that is quite reasonably economical. Such satellites are much too big to build on the ground and launch into space, so they're going to need assembly stations in orbit to put them together. Such stations will, by definition, have a lot of satellite parts on board--spares, in other words. Such satellites are in fact of marginal economic viability for any reasonable (and many unreasonable) reductions in launch costs from Earth, so they very nearly require space-based manufacturing to make sense. In that case, obviously the spare parts are made in space, so it would likely be more costly to fly them down to Earth to supply Earth-based servicing missions. Since such satellites require infrastructure that makes moving people and supplies through considerable delta-Vs cheap and easy, it wouldn't be much trouble to ship them around for repair missions, either. And as far as propellant boil-off...well, use storable propellants. Sure, they're toxic, but that just means you have to isolate them from crew quarters. Such services might also come along with other roles. For example, the spacecraft used to ship crews and parts up to satellites that need repairing might also launch satellites and beyond Earth orbit missions, or the station might also be a propellant depot that stores enough fuel and oxidizer to ship anything to anywhere anyway.

I agree that they are not cost-effective in the first place, and that they would likely have to be built there, but that makes them even less of an analog to KSP (the point of this thread) unless EPL were stock.

So under the assumption that we don't leap so far into the future that we are manufacturing PVs and other items on orbit (well past any foreseeable future, IMO), any resupply will have to come directly from earth. I think if we have to posit delivery of supplies to GEO (lunar?), then manufacturing in GEO, then we might as well assume the station in question is an O'Neill cylinder at L5 a few km in diameter. Then the dv to GEO is only ~1.5 instead of closer to 3-5-4. Heck, you could go more near-future, and posit that such construction would be done by telepresence/robots. Why put people in a dangerous radiation environment for no reason?

Long story short, I don't see GEO repair that requires 24/7/365 space station crew on call for instant repairs (the only reason to leave them there at a station all the time) as being a thing for a long, long time, if ever.

For KSP, we don't have failures, so repair is not even a viable rationale.

Edited September 15, 2016 by tater

[Workable Goblin]

5 hours ago, tater said:

Untrue. Marginal costs are meaningless. Total program cost, divided by launches. Every launch cost around 1.5 billion using that measurement in constant dollars. Had Shuttle not been a thing, HST goes via Titan III? likely (or another variant, this is counterfactual, so they'd make what they needed). The launch is less than 1/2 what the shuttle launch added to HST, and there are not X repair missions at a cost of 1.5B each. If you imagine a more limited lifespan, you take the opportunity cost of not blowing billions a year on Shuttle and have new (better) HSTs popping out every 10 years or so.

No, in this case it's precisely the opposite. Your claim was that it would have made more economic sense to throw away Hubble and build a new one than to repair it. You didn't specify, "In a counterfactual where NASA didn't develop the Space Shuttle and therefore couldn't repair it," so it's totally fair to assume that in either case the Space Shuttle does in fact exist and NASA has in fact spent the money needed to develop it and all the various fixed costs associated with it. In that case, the proper comparison is between the costs that NASA actually did incur as a result of launching the servicing missions--that is, the marginal costs of launching the servicing missions--and the costs of building and launching a new telescope, that is the marginal costs of that option.

Anyway, even if you take the fully-burdened $1.5 billion (in 2011 dollars) cost of launching a shuttle mission to Hubble and add the additional billion to the other costs of the servicing mission, you get $2.2 billion (again, in 2011 dollars). The cost of developing Chandra exclusive the launch was $1.65 billion in 2000 dollars, or $2.15 billion in 2011 dollars (inflated with CPI, which isn't really right but is still better than using 2000 dollars); the cost of James Webb exclusive the launch was estimated to be $8.6 billion (in 2011 dollars); and the cost of Hubble itself exclusive the launch was $4.2 billion (see page 10), also in 2011 dollars. I couldn't find equally reliable costing data for Compton, unfortunately, just some websites saying it cost $617 million without specifying what year or breaking it down into development, launch, and operations. Each of these three observatories also took about twenty years to go from initial proposal to launch (projected launch in the case of James Webb, of course).

So for observatories comparable to Hubble, that is big observatories that require Shuttle or Titan or similarly large launch vehicles, the cost of developing the observatory--that is, everything except launching (and operating)--is, in fact, comparable to the cost of launching a servicing mission, even using the fully burdened launch cost, and can be much larger. In the best case, Chandra, a free launch would have allowed the overall mission to cost about the same amount as one servicing mission, with fully burdened launch costs. In other words, it would still at best cost about the same to build and launch a new telescope as to service Hubble, not less--and with the rather obvious and considerable risk that your new telescope suffers a launch failure and goes precisely nowhere. At least if a servicing mission fails, the telescope is still there and you can keep using it for a time, perhaps long enough to try to service it again.

5 hours ago, tater said:

Bottom line is that it certainly could have been replaced a few times for the multi-billion dollar cost of what, 5-6 total shuttle missions for HST?

In the best case, Chandra-like development costs, maybe (though it would have taken 40 or 50 years to actually launch "a few" Hubble replacements, so it's still questionably advantageous). In the worst case, James Webb-like costs, absolutely not. You'd be paying just as much to build the new telescope as you would have to launch service missions...several service missions, in fact.

5 hours ago, tater said:

So under the assumption that we don't leap so far into the future that we are manufacturing PVs and other items on orbit (well past any foreseeable future, IMO), any resupply will have to come directly from earth. I think if we have to posit delivery of supplies to GEO (lunar?), then manufacturing in GEO, then we might as well assume the station in question is an O'Neill cylinder at L5 a few km in diameter. Then the dv to GEO is only ~1.5 instead of closer to 3-5-4. Heck, you could go more near-future, and posit that such construction would be done by telepresence/robots. Why put people in a dangerous radiation environment for no reason?

Robots can require repairs, too...

5 hours ago, tater said:

For KSP, we don't have failures, so repair is not even a viable rationale.

And thank goodness for that, actually. I used to play BARIS (because, er, KSP didn't exist...), and that has a miserably punishing failure system. I prefer being punished for design errors not random component failures, thank you very much.

[tater]

NASA never uses "not Shuttle" to launch HST in the world of Shuttle (they had to throw everything at it to justify it). So the counterfactual is out the window in that case, so it's either Shuttle, no counterfactual, or yes counterfactual, no shuttle, IMO.

Don't use modern telescope costs as analogs, use telescopes that are---for space telescopes---mass produced. HST was a Keyhole variant, we built several of them. That's why it's it would have been cheaper. The "best case" is not Chandra, the best case is Keyhole 11/12. We built 16 of them (17 counting HST) in ~35 years. About 1 every 8-9 years. In mid-2000 dollars, about 2 billion a pop.

On topic, the rationale needs to include the total cost of the space station as well, if repair is its raison d'etre. ISS was up during the Shuttle program---what if Shuttle had to stop first at ISS, pick up an astronaut, then go to HST? Make sense (still grossly less dv than GEO) then? Fold the whole ISS cost into the repairs? Or another vehicle to get to HST from ISS. Fly costar to ISS, laid on tug, take tug to HST, repair, etc...

Edited September 15, 2016 by tater

[Codraroll]

Can I interrupt the thread a little with some musing?

I just realized that space stations and ground stations in Kerbal Space Program are nearly polar opposites in terms of gameplay:

The game gives you plenty of opportunity to build a wicked cool space station: You send modules up in orbit, you dock them together indefinitely using RCS and docking ports, there are plenty of parts to build with, and they're overall just convenient to build. The entire docking mechanic seems tailored for space station construction. However, the benefit of putting together a space station is marginal at best. You have so few reasons to build them, so all the parts and all the docking is, putting it bluntly, for naught. They're really cool and easy to build, though, so many players build them just for the vanity. They're sort of like epic car jumps in GTA - the game doesn't reward you for it, but it's so fun you do it anyway.

Meanwhile, ground stations. Perfect for flag planting missions, sending Science from the surface of various bodies, giving you a permanent presence outside Kerbin, and of course, you can extract resources from the ground. However, the stock game gives you really few parts to use for your ground station, and the mechanics of assembling them are clunky at best. Lining up the elevation of docking ports is a real pain, even on flat ground, moving the modules in place is cumbersome (can be done with some rover skills, but it's difficult), and to refuel ships with the fuel you extract, you need to land them directly on top of the station. There are many good reasons to build ground bases, but doing so is very difficult and the game lacks a mechanic geared towards it. Ground bases are useful, but so inconvenient to build that many players won't even bother with them.

So overall, you have one type of station with lots of parts, relatively easy construction methods, but no real incentive to build; and another type of station with loads of incentive, but few parts and no convenient construction method.

 

Luckily, mods can make both types of station a lot more bearable. Station Science and its likes gives space stations a legit gameplay purpose. KIS/KAS add fuel lines to make refuelling spacecraft from bases a lot easier. Extraplanetary Launchpads lets you launch craft off other planets with the resources you collect. There are habitat mods that add specific ground station parts, rover mods that make ground station construction more convenient, etc. But still, in stock, it's interesting that the two types of station fill each other out so nicely. One type you can conveniently build without a reason why, one type with all the reasons but none of the convenience.

[tater]

Yeah, the game certainly offers more way to make less useful stations rather than more useful (ISRU) bases.

Stations are basically a circular rationale thing for learning about how to live in space. You build things to live in space so you can learn how to live in space. If life support was added, then station science could become required to unlock new LS capabilities, for example (assuming there was a way to link tech nodes to specific contracts/science.

Here's a loony idea that actually incorporates "crazy contraptions" into KSP in a way that makes sense. Someone had mentioned movies, up thread, right? Add orbital/base construction. Then add contracts to "film" certain nutty contraptions as part of kerbal movies. Like "Stephen Kerman is filming a new blockbuster. He'd like you to haul Stephen, and 4 actors up to Kerbin Alpha (recognizes existing stations), then launch 2 craft containing at least parts X, Y, and Z each, and collide one into the other at at least 30 m/s within 1 km of the station, so it can be filmed."

There's a crazy thing I might actually build.

 

[Workable Goblin]

16 hours ago, tater said:

NASA never uses "not Shuttle" to launch HST in the world of Shuttle (they had to throw everything at it to justify it). So the counterfactual is out the window in that case, so it's either Shuttle, no counterfactual, or yes counterfactual, no shuttle, IMO.

No, there are a series of possible counterfactuals here:

1. NASA builds Shuttle, NASA repairs Hubble. Reality, in other words.
2. NASA builds Shuttle, NASA does not repair Hubble. NASA launches Hubble replacement later, on Shuttle.
3. NASA does not build Shuttle, NASA repairs Hubble.
4. NASA does not build Shuttle, NASA does not repair Hubble. NASA launches Hubble replacement later, on a different LV.

It is perfectly possible that NASA builds Shuttle, launches Hubble on Shuttle, and then doesn't repair Hubble. It would take some subtle work to justify, but it is possible. And going by what you actually said, which is that NASA doesn't repair Hubble, not NASA doesn't build Shuttle, you were talking about (2).

Anyway, as I pointed out the conclusion is the same if you compare (1) to (4) as if you compare (1) to (2); servicing Hubble was not more expensive than building new telescopes.

16 hours ago, tater said:

Don't use modern telescope costs as analogs, use telescopes that are---for space telescopes---mass produced. HST was a Keyhole variant, we built several of them. That's why it's it would have been cheaper. The "best case" is not Chandra, the best case is Keyhole 11/12. We built 16 of them (17 counting HST) in ~35 years. About 1 every 8-9 years. In mid-2000 dollars, about 2 billion a pop.

You cannot extrapolate from spy satellites to astronomical telescopes, they are entirely different beasts even if they happen to share a form factor and have similar optical trains. The key problem is that science marches on. That is, every observatory resolves some astronomical questions and opens up new ones. For example, in the 1970s, when Hubble was designed, no one was considering extrasolar planets and so Hubble was not designed with them as a factor, but in the 1990s and 2000s, when James Webb was, they were one of the hottest topics in astronomy (and still are), so they had a major impact on the telescope's design. To address these new questions, new instruments and possibly different designs are needed, which in turn means that you're not just building a copy of an old telescope with a few tweaks, but a new telescope, meaning that you've gone right back to the beginning. And astronomers are a pretty well-organized and influential political community, for a bunch of scientists, so they're going to push for new telescopes and probably get them (they may regret getting them, but they will probably get them). None of this is true of spy satellites, the basic questions that a spy satellite is supposed to answer (where are the missiles? Where are the tanks? Are the tanks getting ready to roll?) are much more constant than astronomical questions so the only reason to change them up is technical advances. By contrast, every observatory is asking new questions from the one before it, and that almost inevitably means it will have a new design.

For that matter, all kinds of scientific satellites have the same basic issue that makes proposals to "mass produce" them unrealistic and absurd. It's extremely unusual to see serial production of identical scientific satellites in the modern world, with the marginal exception of weather satellites (for the obvious reason that you need more than one to see very much of the world reliably). The only cases tend to involve launch failures (as with OCO2), in which case obviously science can't march on. In any case where it is desirable to include new instruments or alter existing ones, it tends to happen--and the result is that savings are very marginal or non-existent. In the similar (in terms of costs, and scientific control) planetary science program, the Mars 2020 rover, which is a copy of Curiosity with new instruments, is estimated to only save about $400 million relative to the cost of Curiosity, whereas the 1980s Planetary Observer/Mariner Mark II program--an effort by NASA to do precisely what you're suggesting in planetary science by designing a common probe chassis and reusing it in a bunch of missions--was a complete and miserable failure, mostly because a one-size-fits-all chassis fit absolutely nothing and had to be customized to each mission anyway.

Anyway, taking the 2 billion a pop cost and 8-9 year construction time at face value you're...still spending just as much on building new telescopes as you would have repairing Hubble, and you're spending more time on each new telescope than they actually did on the repairs. Yay?

[tater]

HST was designed with a 3m mirror, it was reduced to 2.4---because that's what KH had. The spacecraft was basically the same, obviously the optics (aside from diameter) and instrumentation would be different for astronomical use vs pointing it down. It's not really substantially different as a spacecraft, the form factor chosen was not accidental, it prevented them from having to reinvent the wheel.

Regarding counterfactuals, yes, there are many. My point was that the counterfactual where HST is launched on Titan is necessarily one without Shuttle. I think your others with shuttle are certainly possible for the Shuttle case. I think as long as there is shuttle, HST is launched with shuttle, and is either repaired, not repaired, or replaced/not replaced with shuttle. Mine assumes a world without the money sinkhole that was Shuttle. Repair in such a universe seems pretty unlikely, as well (a sort of "Eyes Turned Skywards" future with LEO-specific Apollo CSMs as manned spacecraft, but they would lack the cargo bay, arm, etc for repairs).

I also did not say mass produced in the generic sense, I said are "for space telescopes" mass produced (bad wording on my part, mea culpa). perhaps that caveat was not clear enough. Any such device is a one off, but for space telescopes "mass" production would likely mean no more than certain commonalities in spacecraft design, not instrumentality. The same is true of KH 11/12. There were built in a few blocks, and even with those distinctions, each was still "one off," and customized vs all the others. Heck, all the Shuttle orbiters were different, though "for space shuttles" they were "mass produced." That phrase was not ideal, but don't lose the forrest for the trees, the point was that such instruments would share common elements that would make making multiples easier.

Thermal regulation, power, pointing, etc, are basically the same between HST and those KH units, changing instruments is easy (look at the COSTAR mission itself). This was not happenstance, the same contractors built them.

The repair missions were certainly dubious in terms of economics. COSTAR was not the only such service mission, there were others at 1.5B$ each for the flight alone. My claim is primarily that HST could have been launched and replaced for less than 2 shuttle missions cost (and I stick to 1.5 B$ per launch, not supposed marginal costs when the shuttle never launched as often as planned).

On the topic of space stations:

As it relates to solar power satellites, you'd need to demonstrate that keeping crew on a station, presumably in LEO, 24/7/365 is economical for power satellite repair. That seems incredibly unlikely. If it was likely to need semi-frequent element repair, it seems like the best solution would be to design a power sat that has a framework built for a sort of "crawler" robot to be able to navigate on, and include spares on the power satellite itself. When a PV element shows sufficient damage to require replacement, the crawler grabs a new element, rides the framework to that section, removes the old unit, and snaps in the new one. They could ship those replacement elements in a container that docks (so they'd send more up as needed), and perhaps it would be possible to have the container deorbit when full of damaged units (the robot delivers the damaged part and places it in the rack where the new part came from). Maybe you could use a solar sail for this (or ion), time doesn't matter, the goal is to slow it until it deorbits.

For satellite repair in general, it's worse, since each spacecraft is likely so unique as to require purpose built parts for any repair. These could certainly be delivered along with regular resupply in advance of a repair, but that means that repairs are on longer planning timeframes, which negates the need to leave the crew on orbit. If it's just a repair tug craft, that could remained docked at such a depot, and the crew brought up as needed... I suppose that requires defining a station such that occasional occupation counts, and it can merely be a parking lot for craft you might use. I tend to think in terms of full-time occupants being required for it to be a "station."

 

Edited September 16, 2016 by tater

[monstah]

20 hours ago, Codraroll said:

They're sort of like epic car jumps in GTA - the game doesn't reward you for it, but it's so fun you do it anyway.

You know, a lot of this game is this, not just space stations.

But, yeah, stations are marginally profitable at best.