How outdated was the Shuttle program?

[Dr Mordrid]

Back as far as 2002 NASA had to buy parts off eBay; 8088 CPU chips, circuit boards, 8" floppy drives etc. in bulk, often scavenging old medical gear. Imagine how it's been since.

NY Times....

[VJ]

This actually does not surprise me... The systems where designed with the hardware of the days in mind, and if it works reliably, why bother to upgrade the hardware? Redesigning the systems will require a lot of work (designing, testing, testing, testing, ...) and may not contribute much.

What does surprise me is that they have no supply chain and have to find parts on the second-hand market. The chips of those days are not complicated, and could easily be made by a chip foundry (in bulk). Similarly, the $500 systemboard mentioned in the article could probably be manufactured quite easily, so I'm surprised they don't have contracts with a chip/board manufacturer to just make the parts they need. Nasa often have contracts with Nikon to supply custom built cameras, so why not for this?

If you look at most commercial airplanes (and especially light aircraft in general aviation), the hardware used is often relatively outdated. The reason there is that certification takes time and testing, and there is little point in investing in new hardware if it gives no benefit. For ultra light aircraft, the certifications are different, so quite often they have rather modern hardware inside, main benefit is that it is lighter.

[TransformX]

On the other hand, consider the kind of leap they made between 8086 and 80386...
16bit computing @ 4.77Mhz vs 32bit computing @ 33Mhz, more than probably drawing considerably less power as well.

[Dr Mordrid]

An advantage in space goes to larger size archetectures - a few ionizations have less potential to disrupt the system. Chips made specifically for satellites also have enhanced error correction. Still, a lot of shuttle support systems are horridly outdated. OTOH modern spacecraft like Dragon, Dream Chaser, CST-100, Prometheus etc. will be using state of the art hardware including multi-redundant avionics, flat panel displays etc.

[UtwigMU]

I remember distinctly how on university we were presented by a medieval illustration depicting how foot measurement unit was derived in Germany. The first 30 or so men to emerge out of church on Sunday were put in line foot behind foot and the average foot was calculated.

On that illustration there was a wooden shed next to church and professor said: If we were to construct a wooden shed today we would do it in same way (regardless that then master builders had rules and proportions handed down through generations and now we have physical formulae and computers).

If the technical problem is the same and laws of physics are the same, a good solution will not change much.

[MultimediaMan]

Some of the discussions I have seen revolve around how to code newer platforms to be more tolerant of hardware anomalies.

From the OS kernel on up there has been some exciting work done on dynamically-adaptable computers (Next Generation Virtualization technologies which integrate UEFI) which can add and remove all "hardware" components and essentially run stateless as far as the hardware is concerned.

Consider: A computer cluster with 3-6 nodes and packing it into a 3U half depth, half rack case. It's not only feasible; I have seen it done...and that's about as far as I can go with that at the moment due to NDA.

[Dr Mordrid]

Know how much a rad tolerant CPU & board runs these days - akin to that on a science mission & using PowerPC tech? A BAE RAD750 like used in Deep Impact, LRO, MRO, Kepler etc. and many satellites runs about $200k per board.

[Elie]

well the same thing can happen now if they design a craft based on core i7 technology, by the time they build and fly missions, they'll be scavenging i7 processors and board from eBay as well
Because in 10 years, who knows what will come out?

[MultimediaMan]

I doubt you will see a space-rated 32nm processor of any kind. Like the Doc says, it is not simply a matter of grabbing an off-the-shelf part and testing the daylights out of it. These things are built specifically to be exposed to high energy particles for extended periods and recover/survive the event. Most desktop processors in the same environment would fault at least once a day, but would recover after a reset...a few times, maybe. Sooner or later a photon will permanently short or divot a circuit trace and then the processor is less useful than the sand it came from.

Hardened processors and older-tech ICs are less apt to be affected because their circuit traces are quite a bit larger; either by virtue of design or the limits of the wafer foundry at the time.

Shielding Electronics in space is an electrical engineering discipline all it's own: and shielding isn't really the right word...it is more like mitigating. Cosmic Rays and related particles have been observed to have blasted through the Moon, the space in between the Earth and the moon, then the Earth itself and keep on going...Friends, that is a LOT of energy, and if a sister satellite and the rock you live on can't stop it, it probably can't be stopped at all. So you build your processors large enough that even a direct hit won't be able to damage anything easily. They've known about the size and composition of Cosmic Rays since the 1950's...they know how to compensate for these things; it is just not very "technically-sweet". Each of the Hubble Telescope's electronics packages are about the size of a large format flatbed scanner: the housing/shielding accounts for over 70% of the mass of the unit; the updated avionics and computer controllers were custom built and certified for 10 years of operation in low earth orbit, and have slightly more processing power than a Pentium 233MMX. Even the laptops the astronauts bring into space tend to be fairly antiquated.

[Dr Mordrid]

A great example of what happens up there is the Oh-My-God particle, a single high speed cosmic ray (actually a proton) detected by the Fly's Eye II instrument operated by the U. of Utah on October 15, 1991. It had the energy of a baseball traveling 55 mph - about 51 Joules.

Researchers at the Pierre Auger Observatory, an instrument spread over a Rhode Island-sized chunk of Argentina, have tracked it and other such high energy particles to the supermassive black holes at the center of galaxies.

[Dr Mordrid]

Now we know the cost/flight of the shuttle program....

Houston Chronicle....

Adding up the final tab for the space shuttle program: $1.5 billion per launch

With just two shuttle flights remaining before NASA retires the fleet we can now begin to close the books on the costs of the space shuttle program.

And that's just what Roger Pielke Jr. and Ray Byerly have done in a letter Nature this week. Among their findings:

• The U.S. Congress and NASA spent more than US$192 billion (in 2010 dollars) on the shuttle from 1971 to 2010.

• The agency launched 131 flights. During the operational years from 1982 to 2010, the average cost per launch was about $1.2 billion.

• Including costs incurred over the life of the program (dating to 1969), this value increases to about $1.5 billion per launch.

Next week the shuttle program will celebrate the 30th anniversary since the first launch, of STS-1. During that time the shuttle has provided innumerable benefits, from learning how to live and work in space to constructing the International Space Station.
>

[SitFlyer]

I guess the shuttle tech is completely worn out? Maybe I'm crazy, but, it sure seems like just about everything is breaking for the LAST few flights, anything to keep people employed for as long as possible???????????????????????????

[Dr Mordrid]

Now you're getting it - to Congress STS is a jobs program, AKA pork, for as many districts as possible. The problem is the system has had severe cost and safety issues since it was in the blueprint stage, and this doesn't even take into account the problems with operations. Poblems.

1) launching 100+ mT of vehicle to place 25mT in orbit at a cost of $1.5B per flight makes no sense when an expendable launcher costs 1/5 that much, or 1/10 that much & 2x the mass once Falcon Heavy flies.

2) ignoring 1, sidesaddle launches of a return vehicle exposes the heat shield to shredded foam, ice, or both. See Columbia. All the CCDev ships launch from the top of the booster where foam & ice shedding are a non issue.

3) even without ice and foam shredding the tiles are fragile enough to fail by other means including vibration - see SRB's which generate thrust oscillations.

4) NO ability to do an abort & save the crew after the SRB's light until they separate 123 seconds into the flight. Before then they're dead. All the CCDev ships will be able to abort from the pad all the way to orbit.

5) NO capability to do a ballistic high-G re-entry if a lifting glide entry fails. Ballistic capability is a fallback with capsules (yes, they generate lift) but totally absent in spaceplanes.

etc. etc. etc.