Block I Apollo Guidance Computer (AGC): How to build one in your basement

by blackstache • 13 years ago

71 10 13 years ago

rglover• 13 years ago

I was thinking about this last night. As a designer (UI/Front-End), I had the alarming thought of "what am I going to do with what I know now in ten years?" I'm fairly intrigued by the idea of space, space travel, and space colonization. But in regard to learning how to design spacecraft instrumentation, I'm at a total loss.

It looks like this will give me a rough rule of thumb for what it's like, despite being deprecated technology.

Does anyone have any resources (books, similar tutorials, etc.) where I can start to dip my toes into this type of work?

ChuckMcM2• 13 years ago

NASA used to publish a journal (NASA Tech Briefs) which included excerpts from papers that they had done. All of the spacecraft human factors work has shown up in various journals. A good place to start is scholar.google.com and look for NASA human factors but any good research library should also have the information.

jk4930• 13 years ago

From your website I assume you're in the US. You could contact the Mars Society. They run analog research missions. Not with spacecraft, but with habitats, rovers, equipment (operated by "astronauts"). You could design and test UIs in the near term and get useful feedback. Plus it's usually a lot less bureaucratic than an official agency or company to have a first look. It's a good entry point. If you see that this works for you, you should use their space industry/agency network to get into further stuff perhaps.

We (oewf.org) have analog missions, too. With rovers, equipment, on-site (analog Mars) control stations. And we have a Mission Control Center, what about better UIs for the ground station? (But I think you should be in Europe for that, working remote is certainly difficult in such a case.)

Another possible partner: Look at the Google Lunar X Prize teams. Some (of the advanced ones) could be interested in UIs for their Mission Control.

yoshokatana• 13 years ago

I'm also into UI stuff, and I found this article really interesting: http://www.projectrho.com/public_html/rocket/controldeck.php

It was linked to in the Star Citizen forums, in a thread that talked about adapting a HOTAS (throttle and joystick system) to be used in space simulators, but it goes into a bit of the UI design of various real-world spacecraft (and the language is heavily influenced by Heinlein, which I kinda love).

eksith• 13 years ago

Your first step might be to visit NASA[1] and take a tour of what astronauts might go through. There's a lot to take in with regard to human psychology in stressful/critical situations. What may seem intuitive and obvious on the ground with your "normal" orientation (which means nothing in 0g) may feel completely different.

Try to walk in their footsteps to see how information should be displayed to them first. Then you can go on what that information might be.

It will be a feedback loop of constant improvement and you will need to tie a lot of that into the backend as well. There was a project called Open Flight Linux[2] which was an effort to duplicate the objectives of Flight Linux[3]. You may want to take a browse at the documentation to see what types of information would be useful to display.

You'll need some sort of existing reference to see what additional data can be pulled up.[4]

Edit: With regard to ChuckMcM's recommendation (which is very useful!) The NASA Tech Briefs should be on your path to glance at what might be in store for you.[5]

  [1] http://www.visitnasa.com
  [2] http://web.archive.org/web/20100306001043/http://openflightlinux.org
  [3] http://web.archive.org/web/20080102134626/http://flightlinux.gsfc.nasa.gov/
  [4] https://www.google.com/search?&hl=en&tbm=isch&q=orbiter+cockpit
      https://www.google.com/search?&hl=en&tbm=isch&q=spaceshiptwo+cockpit
  [5] http://www.techbriefs.com/tech-briefs

Patrick_Devine• 13 years ago

I would start with more terrestrial aircraft instrumentation. They already operate on three axis (pitch, roll, yaw) so tend to be somewhat similar. For commercial stuff, take a look at something like the Garmin G1000.

I started a remake of a space video game a few years back which I didn't end up finishing, but I did tackle a lot of problems like 3D navigation. Here's a screenshot: http://pdevine.users.sonic.net/images/sundog-cockpit.png If you're interested, the code is all there and runs in Python/OpenGL.

33272• 13 years ago

This is absolutely amazing... The minimum amount of computation power, memory involved is just staggering.

Isn't it actually more advanced today to a degree, in terms of getting the job done?

How many lines of code would a moon landing guidance system be today? How many modules and dependencies?

Which one wold be more redundant ( I feel the comparison is obvious here if not: simple program whose redundancy is in its simplicity, and a complex program where redundancy is with multiple levels of checks within)

ChuckMcM2• 13 years ago

Well it depends of course. One of the dependencies is that space is a pretty harsh environment and building systems with high redundancy and checking (especially man rated systems) is pretty important. That said, all of the computation and more that was available in the lunar excursion module can be implemented in triple redundancy within the size and power budget of a mac book air, including 9 degree of freedom IMU.

Dr. Brooks at MIT built a lander system with "modern" (at the time, Pentium-3 level) chips for DARPA and it was pretty straightforward, at a talk I attended he suggested that by this time (actually by 2006) he expected the whole package would be less than $100 worth of parts.

The only 'big' thing left between Earth and a regular visit to the Moon is some sort of on-orbit refuelling capability. Once you knock down that particular challenge lifting the parts into LEO which can then be refuelled and sent to L1, and then refuelled again and sent to the moon and back to L1 to top up, and back to LEO. There are existing systems for all the other pieces [1].

[1] So there are liquid fuelled engines, that can be refuelled under gravity, but none have been tested which can be refuelled in zero G. A proponent at ULA proposed a gas pump that would spin creating the necessary force to move propellant from tank to tank without the need for an additional non-dilutive pressurizing agent. But it is a fair criticism that we don't have zero g refillable engines yet.

33272• 13 years ago

I am reading 'failure is not an option' if you haven't already read it (something tells me you have) highly recommended.

Its sad how the steam of space exploration and development got dispersed after Apollo. If it had continued at that rate who knows where we would be now. Perhaps even a small station at L1 with small crew... Thankfully I feel we might be in the beginning of a new renaissance thanks to lower costs and open source tech. Somehow the $/lb needs to drastically get reduced. There was a space cannon called project harp that could launch 180kg up to 110 mi. Its a shame they stopped it because the costs per kg are insanely cheap. Perhaps with todays robotics components could be accumulated in orbit.

Long story short, we need zero G refueling and zero G assembly. Next gen space vehicles need to be assembled in orbit.

zw123456• 13 years ago

Here is an AGC I did years ago on a Cyclone FPGA,I used a touch LCD for the UI, and it has an IDE for the assembler,. it fits in around 1100 FPGA LE's! https://github.com/donnaware/AGC The interesting thing that I learned in doing this project is how similar present day MCU's are, most are loosely based on the old 8051, which I can see striking similarities.

Block I Apollo Guidance Computer (AGC): How to build one in your basement

Related Stories