Slashdot Mirror

Notice that the comment you are replying to refers explicitly to 1959-1964, while the document you link is talking only about the tail end of that, when things were converging on the part being reliable and off the shelf. If you go back a year or two, such as this document covers via discussing purchase orders for the IC, you'll see things changed rapidly over a couple years. Orders from various companies had to be canceled at times because they could not deliver the part on the time schedule they gave. Some of which were able to come back in a couple months or year after revamping their processes and try again, usually at a cheaper price too. The early orders there show prices on the order of $40 an IC, but they drop to about $10 an IC. The document doesn't discuss much before 1962, but if you go back another year or two, the prices were hundreds of dollars, and originally $1000 an IC. Lead times were months and sometimes more than a year. NASA pushed to companies to get their stuff in order and made some realize they weren't as good at producing the ICs as they thought.

A couple years here makes a big difference, so when you talk about something from 1965 or "the early 60s" that may be missing a huge part of the story. I know some fellow ham hobbyist can go overboard, but I don't know of many that would spring for $7000+ (adjusted price) IC for building a radio or recommend that is something to write up in a magazine about. But if it were And yes, the parts were RTL. Some earlier guidance computer made use of DTL instead, and used less common, more difficult to source combinations of ICs, and presented several problems in their design and production. Other ground computers at the time were starting to testing out the switch to RTL, but not RTL ICs, as it too expensive and not worth the savings in assembly complexity and size.

Well, just like our economists forgetting the economics lessons of the past 80 to 100 years, now NASA is faced with budget cuts after the resulting spectacular economic failures. So, how does NASA respond? By apparently forgetting the engineering lessons of the past 80 to 100 years. Keyword apparently.

I've designed build space electronics, from launch vehicles to earth science instrumentation for low earth orbit, to weather and comm satellites for GEO and even cameras for planetary science missions, and I can tell you with certainty that the iPhone will not operate reliably at orbits much higher than ISS, and even there won't operate for very long. Take them to a very rad hard environment, like Jovian orbit, and they won't function at all. NASA knows this. And I doubt they've forgotten it. (Here is a good introduction on radiation effects, and this is a very good site for diving deep into the topic.)

Rather than assuming that they're idiots, I suspect that either a) they have a very select subset of missions for which they're considering consumer grade mobile phones (e.g., short duration low-inclination LEO missions), or 2) they're intentionally proposing a noncompliant technology for a mission for a reason, such as demonstrating the impossibility of the mission for the proposed price, or perhaps simply in protest of budget cuts.

I'd put my money on 1) but would not be surprised if 2) were true. I've seen it before.

Microcomputers were developed based on needs by NASA to have computers that were light enough to be on a spacecraft because you couldn't fit a room-sized mainframe on an Apollo spacecraft or on the Lunar Excursion Module. So, let's see. We have this little space race thing that ends in the 1970s with NASA pouring money into little teeny solid state computing devices and you get the Apple ][ computer in 1977. And the IBM PC four years later. The last Apollo spacecraft was designed around 1967 more or less so I have to ask the naysayers what they're expecting to see in about ten years now that the ISS is complete. because everybody knows NASA science doesn't contribute to anything down here on earth.

No, they weren't. Microcomputers were developed because of a desire to get the parts count down for electronic desk calculators. That's why the Intel 4004 was built. The Intel 8008 was built to get the parts count down for the Datapoint 2200 terminal, although it was late and Datapoint had to use another approach.

The Apollo Guidance Computer wasn't even close to being a single-chip computer. It was a nice piece of electronics packaging, but it had 5600 gates in 2800 packages. Integration was at the level of a single-chip dual NOR gate. It did the job, but was a technological dead end.

This guy wins the ubergeek crown of all time. He actually built a fully functional version of an AGC out of discrete IC chips. Granted, it is not a true AGC clone, since the exact chips could not be procured in 2000-2004, but he did build a functional workalike of the AGC out of individual TTL chips, and wrote two software emulators, and got his modernized hardware reproduction AGC to run actual original AGC machine code software.

There is one dude who built a working replica of the AGC and let it run the original flight software:

http://klabs.org/history/build_agc/

Lots of good information in those PDFs. Also very interesting that the AGC already ran some simple real-time capable multitasked OS. Simple, but still with a lot of modern ideas in it.

Actually it's the US segment of ISS that uses the 386 based machines (although a couple may go to Pentiums in a few years).

The Orbiter uses the AP-101S, which was also used in military aircraft. NASA has a great deal of published history online regarding Shuttle Avionics here.

Shuttle flight computer uses the 386 double sigma.
here are some other notes:
http://klabs.org/DEI/Processor/shuttle/

, and the summary is available online at:

http://klabs.org/richcontent/Reports/Failure_Reports/as-204/senate_956/index.htm

There was a lot of pressure from some members of congress (particularly Walter Mondale) to shut the program down after the fire. Fortunately, cooler heads prevailed, and the program recovered from the tragedy with the launch of Apollo 7 the next year.

Lamgley,
Paper E3,
Paper 161 and even a 110MB video of students
programming FPGAs at NASA

Lamgley,
Paper E3,
Paper 161 and even a 110MB video of students
programming FPGAs at NASA

David Parnas, a Software Scientist, who formerly served on SDI Committees and who had no moral qualms about death and destricion ended up quitting SDI and debunking it when he realized the whole program wasn't plausible and a huge waste. It still isn't BTW, but politicians don't get science: billions of dollars regularly flushed down the toilet after it.

http://klabs.org/richcontent/software_content/pape rs/parnas_acm_85.pdf
http://www.wordyard.com/2007/01/05/parnas-sdi/feed /

There are a number of source for the information I was quoting. By far the best "original" source of information about the early history of IC production can be found at:

http://technetcast.ddj.com/tnc_catalog.html?item_i d=393

This is a lecture given by none other than Gordon Moore (of Moore's Law fame) that goes into the history of ICs.

There was also some time about (about a year or so, perhaps a little longer) a /. story about the Apollo Guidance Computer and an attempt to completely recreate it as a part of a home simulator. For details about this project, you can look at this website for some much futher information:

http://klabs.org/history/build_agc/

There is other information that I've picked up over the years as well, but these are two sources to get you started if you are really interested in some of the early history of the Apollo mission computers. Of course Wikipedia is also available with some other outstanding information if you look up "Apollo Guidance Computer", particularly if you read through the sources to generate the Wikipedia article.

Thanks! You got to it before I did :) Here's some more items:

1) NASA has known about this at least as early as 1980, *before* the first shuttle flew. The computers onboard the shuttle are IBM AP101Ss with 64K of RAM and capable of a blindingly fast 1.2 million operations per second. Remember overlays from your DOS days? They are used extensively (major mission modes). Every *bit* of RAM is accounted for.

2) I haven't seen any FMEA/FMECA (Failure Modes Effects Criticality Analysis) posts here on Slashdot yet that would justify the change. Even if you decide to go through with it, you are going to be going through about 10 layers of meetings and reviews (or so it would seem). Not to mention the extensive regression testing that needs to occur. This is safety-critical stuff (i.e. you mess up, it doesn't get caught, someone is going to die). You don't make any changes unless you are able to come up with an overriding compelling reason. It's THAT simple.

3) Work-around procedures are in place and have been in place for quite a while. The crew of STS-116 have successfully trained and run the procedures. This is NOT a big deal.

If you want some reading material on the computers onboard the shuttle have a look at the DPS (Data Processing System) section of the Shuttle Crew Operations Manual(highly recommended) as well as NASA publication SP-504 Space Shuttle Avionics System

Let me guess: a defense (space) contractor designed the systems.

That's right. Apollo Abort Planning, p. 25: "Therefore, the sole propulsion source is the SPS because the service module RCS is incapable of performing a burn as large as that required for the TEI maneuver". However, this was recognized, and the SPS had redundancy in almost all components except the engine bell.

There's no rule against reusing old technology like the Apollo Guidance Computer? Or building a low-tech Salvage 1 is there? Granted it's not sexy but it should work. :P

Even better, it's online. A good chunk of the Apollo Guidance Computer documentation (including the assembly source code to Collossus 249, which was the guidance computer program) is available here.

Fascinating stuff.

As you add more states, it becomes less noise resistant.

E.g 2 states, with 0V = 0 and the supply voltage as 1 gives you immunity to noise at less than half the supply voltage. But there are chips that have more than 2 logic levels. MLC flash chips use 4 for example.

http://klabs.org/richcontent/MemoryContent/nvmt_sy mp/nvmts_2002/docs/12/12_dan_p.pdf

And just with normal logic gates you can get extra states, there's a so called "tri state" when the output is disabled for example. You need this so that devices on a bus which are not selected can allow the active devices to talk. But I think the popularity of binary signalling is mostly that it's noise resistant.

Also its good to point out that older equiptment is better in space because of radiation.

Here is a link to a document explaining what happens to objects just floating around earth Space Radiation effects on Low Orbit Objects

Also I should point out that the shuttle does what it was meant to do, why should you change what works?

Once you get some idea of when the sat will be overhead, you can start tuning around the transmit frequency. You have to keep the squelch open, and it helps to have a continuous tuning receiver because it is easier to adjust for doppler shift. I wouldn't think the NASA communications between the ISS and ground control is open for the general public (looks like the ground control systems are not easily heard by a scanner anyway), but the HAM stuff is all just narrowband FM. Good luck.

Regardless of what everyone on slashdot seems to think, rocket engineers know more about rocket engineering than you.

>It only flew once, unmanned. A feat Shuttle can't do, by the way, as it can't land unmanned.
You're confusing unmanned with automated flight. The shuttle certainly contains autoland capabilty, what sense would there in automating everyting *but* the landing, and does the Buran have similar automation (links please)?

As for the USSR's "smart financial call", During the early 1990's, a man rated Buran spacecraft was being prepared for flight, but in 1993 the program was officially terminated. Since your post is wildly inaccurate on all other accounts, I remind you the USSR fell in December 26, 1991.

I did read the failure investigation report (can be found here, search for polar) some time ago and IIRC the most probable failure scenario was a software error involving a single boolean:

MPL was to land under active control (with rocket power, not the air-bag trick). To kill the moter once it had touched down the legs contained contact sensors which were constructed of a pin with a spring, a magnet and a Hall-sensor. The legs were to be extended some time before touchdown.

The problem was the sensors would trigger some intermediate false readings during the leg extension. These false readings toggled a flag, which, once the control system first started looking for contact, immediately killed the engine, having the lander free-fall to death. Clearing the flag after the leg-extension would have saved the mission. The bug was not found because of errors in the software design documents and lack of a system level test. The intermediate false readings were found in a component level test, but its consequences somehow didn't made it in the final design.

I don't want marketing, I want real space travel, and that requires being a little harsh on all the marketing that surrounds this.

How would you define "real space travel"?

Judging by the cockpit view, this sure seems like space travel as far as I'm concerned.

The Wright Brother's big advance was controlled, powered flight. Lots of people could shoot a projectile from one end of the field to the other, which is all (effectively) that was accomplished by Burt Rutan.

SpaceShipOne is equipped with (and makes heavy use of) a reaction control system, which operates in the same general fashion as the reaction control systems on other spacecraft.

plus astronauts would have one sucky time flying the craft by hand to the moon.

For the most part, what the astronauts did was closer to "flying the craft by hand" then the kind of auto-pilot we think of today.

The computers for apollo did pretty straightforward stuff, and were mainly there so the astronauts didn't have to keep doing stuff non-stop. They could still sight stars and calculate there path and manual fire rockets to adjust (like they did in Apollo 13),

The computer did relieve the crew of having to do lots of calculations which allowed them to focus on control of the craft, but some things were done manually - despite the fact that the computer could have done it automatically.

In a transcript (typos in original) of David Scott's remarks, from The Apollo Guidance Computer: A User's View there is an interesting discussion of the auto-pilot capability of the AGC, and then he goes on to say:

"The lunar landing itself could have been done automatically and many times people ask me about that. Could it have been accomplished automatically through the LEM guidance computer? Nobody ever did it. We all felt that when you get that point and you are ging to land on the moon, you have to have your hands on the stick. I like computers and I believe in computers, but it aint going to land me on the moon. I'm going to do that. If something gets screwed up then it is going to me, it isn't going to be the computer. Actually, my thinking at the time was that if a problem did occur it was so time critical that you wouldn't have time take corrective action, so you stay ahead of that problem by flying it manually."

will it run Linux? ... or at least NetBSD?

Yeah, but based on what i read about the original, i bet BillG wishes he could port NT to it :

"Shortly after liftoff of Apollo 12, two lightening bolts struck the aircraft. The current passed through the command module and induced temporary power failure in the fuel cells supplying power to the AGC. During the incident the voltage fail circuits in the computer detected a series of power trenches and triggere several restarts. The computer withstood these without interruption of the mission programs or loss of data.

quote from http://klabs.org/history/history_docs/mit_docs/170 7.pdf

Which kinda redefines "Uptime" for me :)

. . .

This is a link to a a partial tear-down of a Apollo Guidance Computer Logic Unit.

http://klabs.org/mapld04/presentations/session_g/g 1007_hall_s.ppt

on slide three, N.B. the cost : $275,800.00.

now i wonder could the guy in the story have afforded to deal with this as well :

"In the early orbital missions before Apollo, NASA learned that the human animal, confined in a spacecraft for a week or so, was not as clean as might be expected from observations on Earth. This additional constraint had . . far-reaching impact . . All electrical connections and other surfaces had to be corrosive resistant . . . everything had to be hermetically sealed."

eww!

quote from http://klabs.org/history/history_docs/mit_docs/170 7.pdf pages 4-5.

. . .

This is a link to a a partial tear-down of a Apollo Guidance Computer Logic Unit.

http://klabs.org/mapld04/presentations/session_g/g 1007_hall_s.ppt

on slide three, N.B. the cost : $275,800.00.

now i wonder could the guy in the story have afforded to deal with this as well :

"In the early orbital missions before Apollo, NASA learned that the human animal, confined in a spacecraft for a week or so, was not as clean as might be expected from observations on Earth. This additional constraint had . . far-reaching impact . . All electrical connections and other surfaces had to be corrosive resistant . . . everything had to be hermetically sealed."

eww!

quote from http://klabs.org/history/history_docs/mit_docs/170 7.pdf pages 4-5.

Due to popular demand there are now two mirror sites for the AGC Replica project files:

NASA Office of Logic Design

SpaceRef.com

In effect, the [Doppler] shift would push the signal out of synch with the timing scheme used to recover data from the phase-modulated carrier.

My understanding of this sentence was that the phase detection/demodulation works correctly, and "timing scheme" refers to whatever they do to clock the 8192bits/s bitstream into Cassini's onboard computer (the paragraph immediately following the text you quoted also seems to suggest that).

Also, the article states that "the timing scheme was implemented by firmware loaded in Cassini's receiver". There is probably no way one could detect phase changes in a 2060 MHz carrier signal in realtime purely in software.

No, they're not.

The processors in MER are RAD6000's, which are radiation-hardened versions of the RS/6000, the predecessor to the PowerPC (see this for details). The RAD6000's younger brother, the RAD750, is indeed a rad-hardened PowerPC.

As an aside, there is a big difference between a radiation-shielded processor and a radiation-hardened processor. Shielding implies just sticking some kind of rad-absorbent material between the processor and the environment. A rad-hardened processor is actually manufactured in a different way - different gate layout, different design rules, often different materials (Silicon-on-Insulator is popular). These things are done to minimize or prevent the effects of single-event upsets (when a bit is flipped by high-energy particles) and single-event latchups (which basically turn a couple of gates into a glorified short-to-ground). The materials changes may also improve the overall total dose tolerance of the processor. The work required for redesign is one of the reasons that space-qualified rad-hard processors lag the commercial market. The NASA Office of Logic Design has some good papers on space processors available online if you're interested in learning more.

Yes, just as it happened with the moon shot (i.e., the one proposed by Pres. Bush I). Our government of late has a good track record of late of promises that will be met by the next administration.

The space station was a very exciting & challenging idea when the president (Reagan, for those who don't remember) proposed a permanent manned station 'within the decade'. 20 years later and counting, and look what we have.

This guy's really a goofball trying to make the argument against diversity as a tool to gain fault tolerance. NASA makes the argument for diversity in life-critical software systems and NIST studies show it's value in High Assurance Systems. KLabs has found the use of diverse and redundant systems on spacecraft offers high protection against failures due to design deficiencies and that it can offer lower cost where the backup system is used as a lifeboat for the primary system.

A few weeks ago I was at the 2003 MAPLD (Military and Space Applications of Programmable Logic Devices) conference, and one of the talks was by Roger Launius, chief historian for the Smithsonian Air and Space Museum. He talked about the history of NASA, in the context of the Columbia disaster, what he thought lead to the failure, and where NASA could go from here. His outlook was pretty grim, but he made excellent points, which enraged a large portion of those attending the conference, half of whom were NASA employees.

Essentially, he said the Shuttle failed (and he didn't just mean 'crashed', he meant, failed to live up to its hype, to do real scientific work in space, and be cost effective) because it was designed wrong. It was designed to be all things. It was designed to transport people into space. It was designed to transport cargo into space. It was designed to conduct research in space. By trying to do all of these things, it failed to do any of them well. He made a number of other vary salient points about the reasons we should or should not send people into space, and the impact of public opinion and politics.

To keep this OT, I'd have to say, considering the historical perspective I learned from Dr. Launius, I like the capsule approach the best for transporting humans into space. It's cheap, it's effective, and it's less likely to break. I'd like to see NASA design vehicles that are inteded for a specific purpose, and do that purpose well. We have a space station for science that can only be done by humans in space (which there isn't much of...how do you really do microgravity experiments with people on board bumping into stuff, and jarring the place around?), we need a low-cost vehicle for transporting cargo, and a high-safety vehicle for transporting humans.

You mean people like those killed in the widely studied Therac-25 accidents in the late 1980's? Or the US soldiers killed because of a software failure in the Patriot Missle defense system in 1991?

More technical information is found in MAPLD Paper D1 and other reports. NASA Huntsville, NSA, USAF (Eglin), University of South Carolina, George Washington University, George Mason University, San Diego Supercomputer Center, North Carolina A&T and others have StarBridge Hypercomputers they are exploring for diverse applications. The latest StarBridge HC contains Xilinx FPFAs with 6 million gates compared to the earlier HAL-Jr with only 82,000 gates. Costs are nowhere near $26 Million. NASA spent approx 50K for two StarBridge Systems.

You're right on! I was at the JPL-NASA-Smithsonian event this spring with O'Keefe, congressmen, Press, past JPL Directors, etc. which very effectively documented JPL achevements in the past, current research and future missions. I doubt that other NASA Center (with Government restrictions) could accomplish such effective PR. It is outstanding what JPL has achieved for the U.S. public, but as you indicate the public is often confused by who actually does the work and research at NASA as often mission telemetry, video and data on space missions are fed back using the excellent JPL downlink facilities so JPL is often given credit for project actually run by other NASA Centers. For example, how many know that the first successful Mars Landings in 1976 (Bicentenial) were the NASA Langley managed Viking Project? There are actually some that think the '97 Mars Pathfinder & Rover was our first Mars landing.
To review NASA accomplishments, I like Apollo and Solar.

Check out...

G4 PowerPC processor
MIPS R3000 Mongoose
Pentium III/4
RAD6000 variant of IBM's RS/6000

Check out...

G4 PowerPC processor
MIPS R3000 Mongoose
Pentium III/4
RAD6000 variant of IBM's RS/6000

Check out...

G4 PowerPC processor
MIPS R3000 Mongoose
Pentium III/4
RAD6000 variant of IBM's RS/6000

I believe the last chip to achieve a spaceworthy rating was the 486. The Hubble Telescope is currently carrying a 386. :)

http://www.klabs.org/DEI/Processor/386_486/Radiati on/intel.htm