3 ms·

I'm sure the cost of getting the thing there overshadows the cost of construction, but it seems really fabulously overbuilt. I'm sure the martian environment is

by jfoutz 13y ago

I'm sure the cost of getting the thing there overshadows the cost of construction, but it seems really fabulously overbuilt. I'm sure the martian environment is horrible for plastics, but heck, they only need to last 90 days. Let them get brittle and rot. (I'm assuming plastic is lighter and therefore cheaper to transport) I guess it's the difference between doing everything as well as possible and then dealing with the limiting factor, vs doing just enough to meet the 90 day goal. A classic quality vs value trade off.

JimmyL 13y ago
The hard part is defining what "just enough to meet the 90 day goal" is. The rovers were things that had to be built to survive a flight to Mars, go though the atmosphere, land by bouncing on some airbags, and then go drive around in a very hostile environment - which needs a pretty sturdy robot. Over-engineering the robot itself is also a form of insurance. Launching a rocket is hard. Landing a lander with airbags is tricky and very unpredictable. The time from finishing to build a robot and it landing on Mars is a really long time. Funding to pay for all this comes once or twice in a generation. Given all those possible points of failure in things that have nothing to do with the rover itself, why not engineer the rovers to be as resilient as possible to try and make up for some of the possible risks that could happen elsewhere in the project? Put another way - if you're building a race car and the body costs $1000, why not spend $20 (as opposed to $5) on the engine to make sure that of all the things that could break, the engine won't be it.
notdrunkatall 13y ago
Better to overengineer at a cost of an extra ten or twenty million and get many more years out of your rover than to underengineer, have the program fail, and have to face not only the failure to gather the data you wanted, but the political fallout from your failure as well.
digikata 13y ago
If you absolutely want a 99+% chance it would last the first 90 days, then really the chances are that's it's going to last a lot longer. I actually don't think it's a typical quality tradeoff. Because you have to meet a high initial reliability, you end up with higher probability of hitting a endurance beyond your minimal requirement for the device. I'm also guessing w.r.t the structures, you'll have to build to withstand a worst case environment during the launch and landing. Once you've survived those you're "overbuilt" with respect to rolling around the surface, but the strength was needed to survive getting there.
elteto 13y ago
"fabulously overbuilt"... not quite, think about this: you have one chance to launch it into space, send it millions of miles away, reenter atmosphere safely and then survive a hostile environment for only (?!?!?) 90 days! There is no overbuilding, by the nature of the mission itself your Factors of Safety [1] are through the roof!! [1] http://en.wikipedia.org/wiki/Factor_of_safety http://en.wikipedia.org/wiki/Factor_of_safety
- jfoutz 13y ago
  Amusingly, in this case there were two chances. I'm saying if Spirit and Opportunity were constructed less well, perhaps we could have afforded a third with the same budget. I don't really have any honest way to value the difference between 10 years of extra data vs another location (i'd guess the more data is better, but that's hindsight bias) Remember, every ounce of "factor of safety" is tens (or hundreds?) of thousands of dollars.
rcxdude 13y ago
There's also the fact that the cost of design is quite likely a lot higher than the cost of construction. For one-off (or two-off) prototypes it's often worthwhile to overbuild so that the design's easier. Optimising for cost and exact lifetime is expensive.

by jfoutz 13y ago

JimmyL 13y ago

The hard part is defining what "just enough to meet the 90 day goal" is. The rovers were things that had to be built to survive a flight to Mars, go though the atmosphere, land by bouncing on some airbags, and then go drive around in a very hostile environment - which needs a pretty sturdy robot. Over-engineering the robot itself is also a form of insurance. Launching a rocket is hard. Landing a lander with airbags is tricky and very unpredictable. The time from finishing to build a robot and it landing on Mars is a really long time. Funding to pay for all this comes once or twice in a generation. Given all those possible points of failure in things that have nothing to do with the rover itself, why not engineer the rovers to be as resilient as possible to try and make up for some of the possible risks that could happen elsewhere in the project? Put another way - if you're building a race car and the body costs $1000, why not spend $20 (as opposed to $5) on the engine to make sure that of all the things that could break, the engine won't be it.

notdrunkatall 13y ago

Better to overengineer at a cost of an extra ten or twenty million and get many more years out of your rover than to underengineer, have the program fail, and have to face not only the failure to gather the data you wanted, but the political fallout from your failure as well.

digikata 13y ago

If you absolutely want a 99+% chance it would last the first 90 days, then really the chances are that's it's going to last a lot longer. I actually don't think it's a typical quality tradeoff. Because you have to meet a high initial reliability, you end up with higher probability of hitting a endurance beyond your minimal requirement for the device. I'm also guessing w.r.t the structures, you'll have to build to withstand a worst case environment during the launch and landing. Once you've survived those you're "overbuilt" with respect to rolling around the surface, but the strength was needed to survive getting there.

elteto 13y ago

"fabulously overbuilt"... not quite, think about this: you have one chance to launch it into space, send it millions of miles away, reenter atmosphere safely and then survive a hostile environment for only (?!?!?) 90 days! There is no overbuilding, by the nature of the mission itself your Factors of Safety [1] are through the roof!! [1] http://en.wikipedia.org/wiki/Factor_of_safety http://en.wikipedia.org/wiki/Factor_of_safety

jfoutz 13y ago
Amusingly, in this case there were two chances. I'm saying if Spirit and Opportunity were constructed less well, perhaps we could have afforded a third with the same budget. I don't really have any honest way to value the difference between 10 years of extra data vs another location (i'd guess the more data is better, but that's hindsight bias) Remember, every ounce of "factor of safety" is tens (or hundreds?) of thousands of dollars.

rcxdude 13y ago

There's also the fact that the cost of design is quite likely a lot higher than the cost of construction. For one-off (or two-off) prototypes it's often worthwhile to overbuild so that the design's easier. Optimising for cost and exact lifetime is expensive.