Apollo guidance computer, 1969

I watched a 1969 film on YouTube about the inner workings of the Apollo Guidance Computer that helped put a man on the moon. The navigation details are fascinating. I had no idea that the astronauts used a sextant to shoot bearings to stars!

Integrated circuits vintage

7400The film shows logic gates that are packaged in TO cases. (TO cases are small diameter metal cans used to house transistors.) By 1967 these would have been obsolete. In 1967 we were already developing products with readily-available quad NAND gates (using TTL — transistor-transistor logic). These 7400-series parts were packaged in 14-pin plastic DIPs (Dual In-line Packages). The higher spec’d 5400 series were packaged in ceramic DIPs. I’d guess that the Apollo Guidance Computer (AGC) in the film was designed well before that — maybe 1964 or even earlier.

I’m a little confused. The Wikipedia description of the AGC states that it exclusively used Fairchild resistor-transistor logic (RTL) dual NOR gates in a flat-pack. Hmmmm. That’s not what the film shows.

In any case, both the single logic gate in a can and the dual RTL NOR gates in a flat pack would have been obsolete by 1969. I would have thought that NASA projects would have used the latest technologies, not 5 year old technologies. I guess that subsystems within large projects such as Apollo acquire momentum, and once they’ve been proven, the “If it ain’t broke, don’t fix it” motto applies.

It’s an interesting film, regardless.

Advertisements

2 thoughts on “Apollo guidance computer, 1969”

  1. The filming was probably done before 1966. You can learn much more about the design and timeline for the Apollo guidance computer in the 2008 book “Digital Apollo” by David A. Mindell. The MIT Instrumentation Lab (now renamed Draper Labs) evolved this computer (used for both the CM and the LEM) out of the Polaris missile boost guidance system, and the Fairchild RTL technology was selected in 1963, making this one of the first integrated circuit computers in production. These computers consumed 60% of Fairchild’s total integrated circuit output, and drove a MASSIVE improvement in integrated circuit reliability. Reliability has always been NASA’s mission – they are 5 to 50 years behind the technological bleeding edge. Lives and billions of dollars are at stake.

    Those core ropes were the “ROM” for the computer, and were specially created for each mission by hundreds of programmers. Which meant that the code had to be validated, multiple copies the ropes made, the computers assembled around them, and copies of the ropes themselves used in simulators to train the astronauts. Given all the verification steps and lead time, and working backwards from the first manned Apollo 7 missions through multiple flight tests, mockup tests, system integration, etc., all without CAD tools and or integrated software design workstations, they got from program start to the first manned mission in 1968 in a fantastically short time.

    The Shuttle computers were developed much more slowly, and were repackaged IBM System/360s; so a computer architecture first deployed in 1965 was still flying on Shuttle in 2011, 46 years later. Most satellites are still built with mechanical architectures originating in 1950s aircraft. When failure is not an option, neither is the newest technology.

    Like

    1. Thanks very much, Keith, for your informative comment.

      In the summer of 1967 I was first exposed to digital ICs: the SN7400 and SN5400 TTL series. We used them to create a frequency synthesizer which provided 156 to 186 MHz in 100 Hz steps. As I recall, most of the ICs were manufactured by Signetics, National Semiconductor, and Texas Instruments. We tried to use the phenolic plastic-cased 7400s wherever possible, but we specified the ceramic-cased 5400s where noise margins were slim or ambient temperature demanded it.

      The synthesizer had no non-volatile memory. Frequency was selected by a set of 10-position rotary switches with BCD outputs. My first use of RAM was the static RAM 2101 chip. I must admit that there is a visual allure to magnetic core memory, similar to that of vacuum tubes.

      I take your point about NASA wanting to use tried and proven technology. Thanks again for the detail.

      Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s