49

This NASA webpage makes this claim about the Apollo Guidance Computer:

By 1963, MIT - during the testing and development of the AGC Block I units - had ordered and consumed some 60% of the then world's available IC's !

Is this claim plausible -- whether counting already produced ICs or those in production, whether directly by NASA or indirectly through MIT? Please support your answer with references, rather than speculating.

Some information that may help:

  • Jack Kilby made the first integrated circuit (IC) in 1958. The chip did not include the wires between the components.
  • Robert Noyce invented the first monolithic IC -- including the wires on the chip -- in 1959.
  • This was the early "Block I" version of the AGC, which was made of 4100 three-input NOR gates, each packaged as a single gate in a TO-5 can. This was the version that flew on the unmanned flights.
  • Not to be confused with the more famous "Block II" version that flew on the manned flights. That version had 2800 ICs, each a dual 3-input NOR gate in a flatpack.
  • 13
    The claim may be more plausible if it is read as meaning 60% of the world's production of ICs at the time was going to MIT and its AGC project. – Kaz Aug 27 at 19:41
  • 2
    Could you clarify that IC stands for "Integrated Circuit"? It's not something easy to search for and the name has somewhat fallen out of fashion this days. – Diego Sánchez Aug 28 at 20:51
  • There have been a lot more AGCs than those that went on flight (and even there it was always two of them), in case you followed curiousmarks series on youtube about their AGC restoration, it sometimes shows lists of them. – PlasmaHH Aug 29 at 7:31
  • @Nzall Even if it is homework, that would be fine. I'm happy to help with homework. People are supposed to get help with homework if they don't understand it. – only_pro Aug 30 at 16:17
59

Short answer: in the 1960s, NASA was buying and testing large numbers of integrated circuits (most of which would never be used) to make the technology mature.

According to FastCompany:

The MIT Instrumentation Lab tried to design the Apollo computer using transistors, which in the early 1960s were well-settled technology—reliable, understandable, relatively inexpensive. But 15 months into the design effort, it became clear that transistors alone couldn’t give the astronauts the computing power they needed to fly to the Moon. In November 1962, MIT’s engineers got NASA’s permission to use a very new technology: integrated circuits. Computer chips.

and

MIT, on behalf of NASA, bought so many of the early chips that it drove the price down dramatically: from $1,000 a chip in that first order to $15 a chip in 1963, when MIT was ordering lots of 3,000. By 1969, those basic chips cost $1.58 each, except they had significantly more capability, and a lot more reliability, than the 1963 version.

MIT and NASA were able to do all that because for year after year, Apollo was the No. 1 customer for computer chips in the world.

NPR concurs:

The Apollo program didn't invent the microchip, but it guaranteed a huge early market – by 1963, Project Apollo absorbed up to 60 percent of the U.S. supply of chips. The military also installed chips in its Minuteman-II missiles.

Both NASA and the Air Force forced companies like Fairchild Semiconductor to prove the chips' reliability by subjecting them to extreme temperatures and G-forces and rigorous visual and electrical inspections. The result? Apollo helped in part to accelerate the silicon chip revolution, at the pace predicted by Gordon Moore's famous law about the accelerating pace of computing power.

  • 30
    Interesting to note the misperception of Moore's Law as a prediction there, rather than as an observation. Especially since it wasn't written until 1965! – Toby Speight Aug 28 at 8:14
  • 4
    So that is a "No"? NASA bought a lot of US ICs, but not the world (USSR starting that industry domestically 1961), and also not for "the guidance computer" but for developing 'the lot'. – LаngLаngС Aug 28 at 14:48
  • 1
    "The military also installed chips in its Minuteman-II missiles." - there were 1000 of these, starting in 1964. Even if it was MUCH simpler than AGC, it would seem sheer numbers alone would suggest this used more ICs. – Maury Markowitz Aug 28 at 19:31
  • 5
    @LangLangC this got me interested in very early Soviet ICs. Do you have some sources (in Russian is fine) regarding the early mass production of ICs? There's some info on the first experimental ICs in 1961 (in ТРТИ), but everything that I see suggests that actual production (серийное производство) of the early soviet IC models started only in 1964-1965, so it may be plausible that in 1962-1964 USA, being the first, and ordering many thousands of ICs for defence needs, produced more than 90% of worldwide ICs. – Peteris Aug 28 at 20:40
  • 1
    @asgallant - The M-II began development in 1962, was ordered up in 1963, and began arriving in 1964. It entered service in 1965. I think more research is warranted here. – Maury Markowitz Aug 29 at 20:33
21

Edit: as Darrel Hoffman points out, the title question is false; the proper statement is that 60% of the U.S. IC production went to the Apollo Guidance Computer. (Other countries were also producing ICs, e.g. Ferranti's MicroNOR IC (1961) for the British Navy.)

The claim that 60% of US ICs went to the AGC is found in the book Journey to the Moon by Eldon Hall, the inventor of the AGC. He quotes on page 85:

Since building four or five prototypes consumed 60 percent of the country's production of integrated circuits, we can conclude this [i.e. using integrated circuits] was a bold and risky decision.

The reference for the quote is an Apollo program manager at the MIT Instrumentation in an interview in April 1966. Reference: "Interview, Ralph Ragan, MIT, Cambridge MA by Ivan Ertel, April 28, 1966, JSC History Office Transcript."

The same 60% figure appears in the NASA book Computers in Spaceflight: The NASA Experience, again attributed to the Ragan interview:

by the summer of 1963, 60% of the total U.S. output of microcircuits was being used in Apollo prototype construction.

I haven't been able to find the original interview, so I'm not totally happy with these sources.

Page 80 of Journey to the Moon lists the multiple IC purchases for the AGC for 1962 and the first half of 1963. This totals 9866 integrated circuits, many of which were delivered late. In addition, an order for 4100 ICs from Texas Instruments was canceled because they couldn't deliver them. This shows that they ordered a large number of integrated circuits, and companies were having difficulty producing them in these quantities. Incidentally, the price dropped from $115 per chip to $12.75 over the course of these purchases.

The book Economic Impact of Large Public Programs: The NASA Experience page 58 has information on the fraction of ICs used for defense (including but not limited to NASA). The estimates are 100% in 1962, 94% in 1963 and 85% in 1964.

  • 5
    Of course, this cites only the country's production, while the original quote is for the whole world. Granted in the early days, most of them were probably being made in the US, but not all. E.g. the Russians probably needed them for their space program as well, which was ahead of NASA at almost every step until Apollo. – Darrel Hoffman Aug 28 at 13:30
  • @DarrelHoffman did USSR have ICs at all in 1962? The fact they had a need for a particular solution doesn't necessarily mean that they used the same technology to achieve it at the same year as USA did. – Peteris Aug 28 at 20:08
  • 1
    @DarrelHoffman I don't know much about early USSR space guidance systems, but sources such as web.mit.edu/slava/space/introduction.htm suggest that early Soyuz did not have a digital guidance computer because the soviet industry could not yet make one; and that the first such system was Argon 11c computer-museum.ru/english/argon11c.htm which was only started in 1964 and developed in 1968, so it seems plausible that USSR IC production in 1963 was very low, not in thousands. – Peteris Aug 28 at 20:13
  • @Peteris That was just an example - I'm sure there would be other non-spaceflight uses for ICs outside of the U.S., so I'd assume at least somebody else was making them. Whether it's enough to make a dent in the statistics for U.S. vs. worldwide production I couldn't say... – Darrel Hoffman Aug 28 at 20:20
  • 2
    @DarrelHoffman quite on the contrary, if we know that a particular thing was invented just now, initially only the inventors of the process have the ability to make it (and barely so - e.g. Texas Instruments being unable to deliver), and they're used 100% for defense applications so the know how is presumably not available to other countries, then we'd expect years before somebody else would be able to make ICs at scale. The fact that USSR space program wanted digital control computers but didn't get them for years is some evidence that it wasn't simply possible to buy them from "somebody". – Peteris Aug 28 at 20:30
6

To add to the already good answers: according to the book Digital Apollo (which also cites the 60% number), that AGC consumed such a large percent of the world market was intentional and strategic decision on NASA's part, as a way to both influence the design of the ICs they bought and to ensure that the supplier would stay in business:

Only one company, Fairchild, made the required chips. Would they remain available for the entire length of the project, or were they just another passing fad in electronics?... Would Apollo be a sufficiently large customer to encourage the suppliers to stay in the business? To increase the demand, [Eldon] Hall proposed all the ground test equipment for the computer be build with the MicroLogic circuits as well.

(page 126)

The major guidance components, such as the inertial unit and the accelerometers, were supplied under specific subcontracts. The integrated circuits, which proved as critical as the larger components, were simply bought from vendors... The only leverage Hall had over the suppliers, then, was as a customer. He hoped that by buying large numbers of a single part he could persuade them to improve their processes.

3

This appears to be quite plausible because until 1964 integrated circuits were expensive and unpopular. Leslie Berlin covers the early history of Fairchild in ICs in The Man Behind the Microchip: Robert Noyce and the Invention of Silicon Valley, from which the following quotes are taken.

Fairchild introduced Micrologic in 1961 and attracted a lot of interest. However, despite refusing to license their patents in order to limit competition:

The reaction was gratifying but did not translate into widespread adoption. By the end of 1961, Fairchild had sold fewer than $500,000 of its Micrologic devices, which were priced at about $100 apiece. Texas Instruments, the only other major supplier, was having such problems selling integrated circuits that it cut prices from $435 to $76 in 90 days. The move had little effect.

Customers objected to them because they were expensive, possibly less efficient in certain ways, and because they might limit their own ability to design how they wanted:

If the integrated circuit manufacturers designed and built the circuits themselves, what would the engineers at the customer companies do? Moreover, why would a design engineer with a quarter-century’s experience want to buy circuits designed by 30-year-old employee of a semiconductor manufacturing firm? And furthermore, while silicon was ideal for transistors, there were better materials for making the resistors and capacitors that would be built into the integrated circuit. Making these other components out of silicon might degrade the overall performance of the circuits.

As late as the spring of 1963, most manufacturers believed that integrated circuits would not be commercially viable for some time, telling visitors to their booths at an industry trade show that “these items would remain on the R& D level until a breakthrough occurs in technology and until designs are vastly perfected.

As other answers have mentioned, there were also plenty of production issues in the early '60s:

A brief flurry of customer interest in 1962 proved little more than a frustration because Fairchild could not build circuits in any real quantity. “Inventory on all [Micrologic] short,” Noyce wrote with great irritation in May 1962, “13K [orders] backlog.” He worried that Signetics, the company started by the former Fairchild integrated circuit team, might be first in line to meet customer demand if it ever materialized in any serious way.

So between the production problems and the lack of customer intrest, through 1962 integrated circuits were selling only to the government and military, and not in huge quanitites:

By the end of 1962, Noyce had to admit that the integrated circuit had thus far had “less than a 10 percent effect on our conventional sales.” He called his staff together on a weekend to discuss “how to get more effort on micro ckts [circuits].”

Eventually they would have to start pushing ICs toward commercial customers, but this took a couple of years:

The need to crack the commercial market became more acute after 1962, when Defense Secretary Robert McNamara instituted changes in military procurement and cost-cutting measures that began shrinking the defense market for integrated circuits. (The military would move from buying 100 percent of integrated circuits produced in 1962 to 55 percent of those made in 1965.)

It was in the spring of 1964 that they finally cracked (or created) the commerical market for ICs:

To [Noyce] it was obvious that despite their other purported concerns about the integrated circuit, customers’ primary objection to the new technology was its cost.... If their technical objections had been met and they still were not buying, the problem had to be the price tag. Accordingly, in the spring of 1964, Noyce made a little-discussed but absolutely critical decision. Fairchild would sell its low-end flip-flop integrated circuits for less than it would cost a customer to buy the individual components and connect them himself, and less than it was costing Fairchild to build the device.

...

This brazen play for market share leapfrogged Fairchild to the top position in number of circuits sold each year. At the same time, the entire market for integrated circuits took off. In early 1964, industry-wide integrated circuit sales for 1966 were projected to reach $58 million, or 8 million units at an average cost of $7.25 a unit. Within a year of the Fairchild-triggered price drop, estimates had been upped more than 150 percent, to $157 million annually— remarkable growth, given that the average price per unit had dropped.

Less than a year after the dramatic price cuts, the market had so expanded that Fairchild received a single order (for half-a-million circuits) that was equivalent to 20 percent of the entire industry’s output of circuits for the previous year. One year later, in 1966, computer manufacturer Burroughs placed an order with Fairchild for 20 million integrated circuits.

So to summarize, until early 1964 it would be easy for any large customer for ICs to be buying a substantial fraction of total production because the market was just so darn small.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.