8 ms·

A big part of the cost of space missions is not making things as light as possible, but developing the scientific instruments. More upmass doesn't always make d

by gene-h 5y ago

A big part of the cost of space missions is not making things as light as possible, but developing the scientific instruments. More upmass doesn't always make designing sensitive scientific instruments easier. If said scientific instruments need to go on a rover, mass still matters. More mass can limit rover mobility and at some point the rover requires more power than can be provided with RTGs or solar cells.

qayxc 5y ago
Exactly. Way too many people seem to ignore the cost of developing and certifying scientific instruments. There's nothing gained from sending 50t of metal to the outer solar system. Scientific instruments are usually purpose-built and while lifting mass restriction can bring down cost, you'd still have to account for redundancy and precision because no one's going to fly out to Jupiter to properly unfold that darn antenna.
- neolefty 5y ago
  Nuclear power sources are still mighty expensive too. Solar is hard past Mars, and near impossible past Jupiter, AFAIK.
- cookingmyserver 5y ago
  > Way too many people seem to ignore the cost of developing and certifying scientific instruments. It gets ignored because it is a symptom of not being able to cheaply launch probes. You spend all of this time researching and developing a scientific instrument only to make two of them. One to launch the other to go on a diagnostic model (maybe a third for a launch backup). You spread out the cost of development over more units and bring down the per unit cost of manufacturing (although probably not that much given we are still talking quantities in the tens or at most hundreds of units). But without a cheap launch system there is a fixed cost per unit of hundreds of millions to get it into space, so there is very little point in producing more than what you need for just that mission. Of course certain missions are going to require special specs, but I do think that generic, somewhat modular spaceprobes are around the corner now that launch costs are going down. If you are going to retire a space craft that you've made a profit off of shuttling cargo to LEO might as well discard of it by sending it on one last mission into deep space with a couple of scientific missions onboard. We will end up spending more money overall, but cost per unit of science (however you want to measure that) will go down.
  - qayxc 5y ago
    > It gets ignored because it is a symptom of not being able to cheaply launch probes. Is it, though? Let's do a quick and very superficial analysis of the costs here. Say we want to send a probe the outer solar system, like the moons of Jupiter or the Saturn system. There are two unique challenges that such probes are faced with: • on Jupiter, the magnetic field of the planet requires hardened electronics and shielding • in the Saturn system and beyond, nuclear power is the only option for powering your instruments Neither of these points are in any way shape or form correlated with launch costs. In fact, in order to get RTGs you need to build, maintain and operate an entire specialised nuclear facility. Alternatively you could use small nuclear reactors, but that may or may not interfere with certain instruments. The cost of hardening electronics and shielding is also not going away even if launch costs were non-existent. It also doesn't matter much whether you produce 10 or 1000 of such specialised processors - the cost will still be orders of magnitude higher than comparable commodity options. It's not getting any better if we turn inwards instead and consider Venus or Mercury. Both again pose incredible and unique engineering challenges as do most scientifically interesting targets in the solar system. How exactly would a cheaper launch cost solve the problem of a Venus sample-return mission? Is it the expensive launch that prevents us from sending an orbiter out to Pluto? How does more and cheaper payload help with getting a probe into the oceans of Europa, Ganymede, or Enceladus? I could go on, but I think you get what I mean. edit: there's also the human factor that I conveniently left out - every mission requires a staff of people (both scientists and engineers) for the entire mission duration; that cost is also unrelated to launch costs
    - cookingmyserver 5y ago
      > The cost of hardening electronics and shielding is also not going away even if launch costs were non-existent. Of course, but what about hardened electronics exempts it from the principles I discussed with scientific instruments in general? > the cost will still be orders of magnitude higher than comparable commodity options. Irrelevant, no one is comparing the cost to commodity hardware, we are comparing it to what the cost would be if we don't increase the number of scientific missions and continue with the current demand. > in order to get RTGs you need to build, maintain and operate an entire specialized nuclear facility. Yes, needing more material is going to cost more money, but does the cost per unit stay the same or go up the more you buy? We are going to need to increase production Pu used in RTGs anyways, if we are going to build facilities to might as well utilize them to the max. > How exactly would a cheaper launch cost solve the problem of a Venus sample-return mission? There will be unique engineering challenges that require unique solutions, but engineers will have a large shelf of parts to choose from if they have a specific issue they don't have to invent the wheel for. > that cost is also unrelated to launch costs That one is true. Missions that involve complex maneuvers, landings, driving rovers, etc, will require large amounts of staff and engineers. Other missions that don't involve those things could utilize a shared staff, so something like a general purpose probe network taking measurements across the solar system. Fleet management tools, policy, and procedures need to be developed so we can manage assets like we do with other LEO constellations.
vanattab 5y ago
This. The JWST cost/will cost about 10 billion dollars. The launch costs are negligible.
- TeMPOraL 5y ago
  The JWST costs that much primarily because it's a one-off, which is a consequence of high launch costs. If you drop the launch costs by 100x, this means being able to launch 100x JWSTs for the same price, meaning it makes sense to optimize for production costs instead of reliability - as if one of your telescopes fail, you can always launch/use another. So now it makes sense to focus on using mass-manufactured parts instead of specialty ones, which drops the production costs significantly. It might make sense to design a telescope platform instead, opening more opportunities to exploit economies of scale.
  - giantrobot 5y ago
    > If you drop the launch costs by 100x, this means being able to launch 100x JWSTs for the same price As the GP pointed out the launch cost of the JWST is a small fraction of the program's overall cost. Launching a hundred JWSTs would mean building a hundred JWSTs and doing so is not cheap. Even with economies of scale producing various components for the hundred JWSTs those all need to be tested and verified. That's an expensive process that doesn't really benefit from economies of scale. For an instrument the size of JWST you can't just spam launches with the assumption some will fail. It's a giant instrument. A catastrophic failure would see it land mostly intact in a populated area or fill a huge orbit with dangerous debris. Even if it was just non-operational it would not occupy an orbital slot that's unusable by other instruments. Launch costs are almost always a small portion of overall mission costs. Just because launch costs go down doesn't mean space missions can magically scale up. The goal is to launch useful scientific/commercial instruments or space vehicles. We don't just launch hunks of lead into orbit for funsies. Lower launch costs are nice but they're not going to magically make all space missions cheap or easy.
    - TeMPOraL 5y ago
      > Even with economies of scale producing various components for the hundred JWSTs those all need to be tested and verified. That's an expensive process that doesn't really benefit from economies of scale. My point is that all those processes are expensive because the final result has to work the first time, because launches are expensive and infrequent. Drop launch costs enough, you don't need so much reliability, so everything across the board suddenly becomes much cheaper. Which then opens many further opportunities to reduce costs.

by gene-h 5y ago

qayxc 5y ago

Exactly. Way too many people seem to ignore the cost of developing and certifying scientific instruments. There's nothing gained from sending 50t of metal to the outer solar system. Scientific instruments are usually purpose-built and while lifting mass restriction can bring down cost, you'd still have to account for redundancy and precision because no one's going to fly out to Jupiter to properly unfold that darn antenna.

neolefty 5y ago
Nuclear power sources are still mighty expensive too. Solar is hard past Mars, and near impossible past Jupiter, AFAIK.
cookingmyserver 5y ago
> Way too many people seem to ignore the cost of developing and certifying scientific instruments. It gets ignored because it is a symptom of not being able to cheaply launch probes. You spend all of this time researching and developing a scientific instrument only to make two of them. One to launch the other to go on a diagnostic model (maybe a third for a launch backup). You spread out the cost of development over more units and bring down the per unit cost of manufacturing (although probably not that much given we are still talking quantities in the tens or at most hundreds of units). But without a cheap launch system there is a fixed cost per unit of hundreds of millions to get it into space, so there is very little point in producing more than what you need for just that mission. Of course certain missions are going to require special specs, but I do think that generic, somewhat modular spaceprobes are around the corner now that launch costs are going down. If you are going to retire a space craft that you've made a profit off of shuttling cargo to LEO might as well discard of it by sending it on one last mission into deep space with a couple of scientific missions onboard. We will end up spending more money overall, but cost per unit of science (however you want to measure that) will go down.
- qayxc 5y ago
  > It gets ignored because it is a symptom of not being able to cheaply launch probes. Is it, though? Let's do a quick and very superficial analysis of the costs here. Say we want to send a probe the outer solar system, like the moons of Jupiter or the Saturn system. There are two unique challenges that such probes are faced with: • on Jupiter, the magnetic field of the planet requires hardened electronics and shielding • in the Saturn system and beyond, nuclear power is the only option for powering your instruments Neither of these points are in any way shape or form correlated with launch costs. In fact, in order to get RTGs you need to build, maintain and operate an entire specialised nuclear facility. Alternatively you could use small nuclear reactors, but that may or may not interfere with certain instruments. The cost of hardening electronics and shielding is also not going away even if launch costs were non-existent. It also doesn't matter much whether you produce 10 or 1000 of such specialised processors - the cost will still be orders of magnitude higher than comparable commodity options. It's not getting any better if we turn inwards instead and consider Venus or Mercury. Both again pose incredible and unique engineering challenges as do most scientifically interesting targets in the solar system. How exactly would a cheaper launch cost solve the problem of a Venus sample-return mission? Is it the expensive launch that prevents us from sending an orbiter out to Pluto? How does more and cheaper payload help with getting a probe into the oceans of Europa, Ganymede, or Enceladus? I could go on, but I think you get what I mean. edit: there's also the human factor that I conveniently left out - every mission requires a staff of people (both scientists and engineers) for the entire mission duration; that cost is also unrelated to launch costs
  - cookingmyserver 5y ago
    > The cost of hardening electronics and shielding is also not going away even if launch costs were non-existent. Of course, but what about hardened electronics exempts it from the principles I discussed with scientific instruments in general? > the cost will still be orders of magnitude higher than comparable commodity options. Irrelevant, no one is comparing the cost to commodity hardware, we are comparing it to what the cost would be if we don't increase the number of scientific missions and continue with the current demand. > in order to get RTGs you need to build, maintain and operate an entire specialized nuclear facility. Yes, needing more material is going to cost more money, but does the cost per unit stay the same or go up the more you buy? We are going to need to increase production Pu used in RTGs anyways, if we are going to build facilities to might as well utilize them to the max. > How exactly would a cheaper launch cost solve the problem of a Venus sample-return mission? There will be unique engineering challenges that require unique solutions, but engineers will have a large shelf of parts to choose from if they have a specific issue they don't have to invent the wheel for. > that cost is also unrelated to launch costs That one is true. Missions that involve complex maneuvers, landings, driving rovers, etc, will require large amounts of staff and engineers. Other missions that don't involve those things could utilize a shared staff, so something like a general purpose probe network taking measurements across the solar system. Fleet management tools, policy, and procedures need to be developed so we can manage assets like we do with other LEO constellations.

vanattab 5y ago

This. The JWST cost/will cost about 10 billion dollars. The launch costs are negligible.

TeMPOraL 5y ago
The JWST costs that much primarily because it's a one-off, which is a consequence of high launch costs. If you drop the launch costs by 100x, this means being able to launch 100x JWSTs for the same price, meaning it makes sense to optimize for production costs instead of reliability - as if one of your telescopes fail, you can always launch/use another. So now it makes sense to focus on using mass-manufactured parts instead of specialty ones, which drops the production costs significantly. It might make sense to design a telescope platform instead, opening more opportunities to exploit economies of scale.
- giantrobot 5y ago
  > If you drop the launch costs by 100x, this means being able to launch 100x JWSTs for the same price As the GP pointed out the launch cost of the JWST is a small fraction of the program's overall cost. Launching a hundred JWSTs would mean building a hundred JWSTs and doing so is not cheap. Even with economies of scale producing various components for the hundred JWSTs those all need to be tested and verified. That's an expensive process that doesn't really benefit from economies of scale. For an instrument the size of JWST you can't just spam launches with the assumption some will fail. It's a giant instrument. A catastrophic failure would see it land mostly intact in a populated area or fill a huge orbit with dangerous debris. Even if it was just non-operational it would not occupy an orbital slot that's unusable by other instruments. Launch costs are almost always a small portion of overall mission costs. Just because launch costs go down doesn't mean space missions can magically scale up. The goal is to launch useful scientific/commercial instruments or space vehicles. We don't just launch hunks of lead into orbit for funsies. Lower launch costs are nice but they're not going to magically make all space missions cheap or easy.
  - TeMPOraL 5y ago
    > Even with economies of scale producing various components for the hundred JWSTs those all need to be tested and verified. That's an expensive process that doesn't really benefit from economies of scale. My point is that all those processes are expensive because the final result has to work the first time, because launches are expensive and infrequent. Drop launch costs enough, you don't need so much reliability, so everything across the board suddenly becomes much cheaper. Which then opens many further opportunities to reduce costs.