Great article. I'd never thought about a spacecraft ADI having a third axis. Sadly, a nitpick - Bill Lear's F-5 autopilot was not, as far as I can tell, in any way connected to the Northrop F-5 fighter jet.
Thanks. You are correct about the F-5 autopilot, so I fixed that. It turns out that it was used in planes such as the C-47, C-60, C-45, and B-26, but is unrelated to the F-5.
>The Command Module for Apollo used a completely different FDAI (flight director-attitude indicator) that was built by Honeywell.
That's surprising. Was there any requirement that necessitated them to be different parts, or it's just because different suppliers were chosen by Grumman/North American?
It's probably a combination of different suppliers being chosen, and everyone wanted a piece of the pie. But it's annoying when I figure out how something works in the Lunar Module and then discover that the Command Module is completely different. Not to mention that the Saturn V is a whole different world.
Yes, in the movie, Lovell says "What's the frappin' attitude?" as the 8-ball rolls out of control. The actual Apollo 13 transcript has nothing like that, interestingly enough.
Cannon is famous for their metal-shell circular connectors to the point that it's like Kleenex. "Everyone" knows what a "Cannon connector" looks like, so I was not expecting a D-sub form factor!
The Soviet Globus is similar in some ways, but also has some major differences. As you mentioned, the ball shows the spacecraft's position over the earth, rather than showing the spacecraft's orientation in space, so the ball looks like a globe with landmasses and everything. The ball rotates along two axes, not three. Moreover, the Globus doesn't have any external inputs; it rotates the ball according to a preset track, regardless of where you actually are.
That's a 'kunst' of UI (a gem?). One look and you instantly know the orientation of your craft.
As an amateur astro-pilot (1000h in KSP and 200+ in Flight of Nova, both flight simulators with realistic orbital mechanics) I'd like to say that in modern cockpit of the fusion propelled ships in FoA, the one thing I'm missing from Apollo-style flight instruments of KSP is the Nav-Ball.
The jet-fighter-like "ladder" style attitude meter can't be read with just one look. You need to focus to see the numbers next to the ladder steps. And then another look at the compass for a full reading. 3s of focus (away from controlling the ship) vs. 0.5 (that your subconscious has most likely already interialized).
To put that 3s into perspective, according to ship readings, Apollo 11 had <20s fuel left when it touched down on the moon.
There are many different Shuttle simulators. The simulator photo in my post is one of the Shuttle Mission Simulators (SMS), now at Stafford Museum in Oklahoma. The Shuttle Avionics Integration Laboratory (SAIL) is a different simulator for avionics testing (rather than astronaut training) and is currently in Houston.
I wouldn't worry too much. There's some clever mechanical stuff in there, but I don't see anything that a competent engineering team couldn't produce in 2025.
The comment about their Manufacturing dept not wanting to deal with it, leaving the Engineering group to build the devices was amusing, but not altogether surprising.
Side: The space age progress of the 1950s and 1960s, is one of the most baffling levels of technological progress. Analog computing (as the sibling comment here by jschveibinz points out about OP article's tech) (and very primitive digital computing), fuel cells, advanced rocketry without FEA/CFD simulation software, literally de-accelerating to land on the moon, and then thrusting to take off again, the list goes on.
kens - Are the collectors of the output transistors on the amplifier boards connected to the metal can? I can see from the photo that the heatsinks don't touch (there's a gap between them for the capacitors). Did they use nylon screws to prevent an electrical path through the frame?
For TO-5 bipolars, it was common for the collector to be connected to the case. I wouldn't say that's universally true but I don't recall any exceptions off the top of my head.
I think it's "generally" true. I saw a mention of some where the transistor had 4 pins: Emitter, collector, base, and ground (can). Presumably for RF shielding and they wanted a higher quality connection to the can without soldering/clamping it directly.
Vibe coders? No. But I can easily see the electrical and mechanical engineers that I've worked with building modern versions of this. The software guys will laugh and say, "give me a 3-axis rate gyro and accelerometer and we can build it with no moving parts."
And I say that as a software guy who has a handful of resolvers in a drawer somewhere :-)
I think people radically underestimate how much capability we have. It's rarely necessary to do things exactly the way they were done 60 years ago.
Its always hard to talk about things or people „in general“. Sure, there are some smart people now and there wwre lots of not so smart people back then. This was not a statement that needed to be disproved. This was a statement about engineering culture in general
The strong impression I always get from the entire Apollo program is "they didn't know it couldn't be done at the level of technology available, so they did it anyway".
And some guys who had previously succeeded at solving difficult problems under time pressure with limited money. Well, the money was more limited on one side of the war.
> 3. The FDAI's signals are more complicated than I described above. Among
> other things, the IMU's gimbal angles use a different coordinate system from
> the FDAI, so an electromechanical unit called GASTA (Gimbal Angle Sequence
> Transformation Assembly) used resolvers and motors to convert the
> coordinates.
I'm willing to bet that these days, that transformation would be done in software! If you have to drive the ball, you'll still need the motors, and the resolvers for feedback, but the coordinate system transformation is "just math."
This really isn't that different than many software tasks, just a different set of basic tools and jargon. When non computer people read a Haskell article that's how it sounds to them.
Author here for your Apollo questions :-)
Great article. I'd never thought about a spacecraft ADI having a third axis. Sadly, a nitpick - Bill Lear's F-5 autopilot was not, as far as I can tell, in any way connected to the Northrop F-5 fighter jet.
Thanks. You are correct about the F-5 autopilot, so I fixed that. It turns out that it was used in planes such as the C-47, C-60, C-45, and B-26, but is unrelated to the F-5.
>The Command Module for Apollo used a completely different FDAI (flight director-attitude indicator) that was built by Honeywell.
That's surprising. Was there any requirement that necessitated them to be different parts, or it's just because different suppliers were chosen by Grumman/North American?
It's probably a combination of different suppliers being chosen, and everyone wanted a piece of the pie. But it's annoying when I figure out how something works in the Lunar Module and then discover that the Command Module is completely different. Not to mention that the Saturn V is a whole different world.
I mainly remember this because he refers to it as the 'frappin 8 ball' in the Apollo 13 movie, if my memory serves.
Yes, in the movie, Lovell says "What's the frappin' attitude?" as the 8-ball rolls out of control. The actual Apollo 13 transcript has nothing like that, interestingly enough.
Links: https://archive.org/details/apollo1319959231994/page/n92/mod... https://www.nasa.gov/wp-content/uploads/static/history/alsj/...
same here, he sure does
I remember a similar thing from the, IIRC, F-104.
Back in the day, this would be have been a good homework assignment for an EE analog controls class.
Great article!!!
I'm kind of a connector geek. Never heard of the MDB1 type before, here's a pic: https://www.digikey.com/en/products/detail/itt-cannon-llc/MD...
Cannon is famous for their metal-shell circular connectors to the point that it's like Kleenex. "Everyone" knows what a "Cannon connector" looks like, so I was not expecting a D-sub form factor!
We had an article on HN last year about a similar Soviet era device. It was a globe that showed the position of the spacecraft relative to the earth.
The Soviet Globus is similar in some ways, but also has some major differences. As you mentioned, the ball shows the spacecraft's position over the earth, rather than showing the spacecraft's orientation in space, so the ball looks like a globe with landmasses and everything. The ball rotates along two axes, not three. Moreover, the Globus doesn't have any external inputs; it rotates the ball according to a preset track, regardless of where you actually are.
My three articles on the Globus had the following HN discussions: https://news.ycombinator.com/item?id=34468212 https://news.ycombinator.com/item?id=35311300 https://news.ycombinator.com/item?id=35038710
That's a 'kunst' of UI (a gem?). One look and you instantly know the orientation of your craft.
As an amateur astro-pilot (1000h in KSP and 200+ in Flight of Nova, both flight simulators with realistic orbital mechanics) I'd like to say that in modern cockpit of the fusion propelled ships in FoA, the one thing I'm missing from Apollo-style flight instruments of KSP is the Nav-Ball.
The jet-fighter-like "ladder" style attitude meter can't be read with just one look. You need to focus to see the numbers next to the ladder steps. And then another look at the compass for a full reading. 3s of focus (away from controlling the ship) vs. 0.5 (that your subconscious has most likely already interialized).
To put that 3s into perspective, according to ship readings, Apollo 11 had <20s fuel left when it touched down on the moon.
This was actually mentioned in a recent talk by Freya Holmér --- I believe this one:
https://www.youtube.com/watch?v=hUlvxaQBW78
I wonder if that simulator was OV-095 at SAIL.
https://spaceflightblunders.wordpress.com/2017/03/31/ov-095-...
EDIT: Ah. It almost certainly was:
https://www.superstock.com/asset/oct-astronauts-frederick-ri...
There are many different Shuttle simulators. The simulator photo in my post is one of the Shuttle Mission Simulators (SMS), now at Stafford Museum in Oklahoma. The Shuttle Avionics Integration Laboratory (SAIL) is a different simulator for avionics testing (rather than astronaut training) and is currently in Houston.
Ken once again proves he's one of the greatest publishers on Hacker News.
Thank you so much for this article. We.read about all the amazing technology that was created for Apollo but this explains one in detail.
I worry with all the outsourcing over the past few decades that these and even basic engineering manufacturing technologies are being lost.
I wouldn't worry too much. There's some clever mechanical stuff in there, but I don't see anything that a competent engineering team couldn't produce in 2025.
The comment about their Manufacturing dept not wanting to deal with it, leaving the Engineering group to build the devices was amusing, but not altogether surprising.
Side: The space age progress of the 1950s and 1960s, is one of the most baffling levels of technological progress. Analog computing (as the sibling comment here by jschveibinz points out about OP article's tech) (and very primitive digital computing), fuel cells, advanced rocketry without FEA/CFD simulation software, literally de-accelerating to land on the moon, and then thrusting to take off again, the list goes on.
Brings me back to playing Kerbal Space Program
kens - Are the collectors of the output transistors on the amplifier boards connected to the metal can? I can see from the photo that the heatsinks don't touch (there's a gap between them for the capacitors). Did they use nylon screws to prevent an electrical path through the frame?
For TO-5 bipolars, it was common for the collector to be connected to the case. I wouldn't say that's universally true but I don't recall any exceptions off the top of my head.
I think it's "generally" true. I saw a mention of some where the transistor had 4 pins: Emitter, collector, base, and ground (can). Presumably for RF shielding and they wanted a higher quality connection to the can without soldering/clamping it directly.
Unfortunately, I don't have the FDAI handy to check this.
When I see something like this my first thought is: „there is absolutely no way current gen vibe coders and engineers will be able to replicate this“
Vibe coders? No. But I can easily see the electrical and mechanical engineers that I've worked with building modern versions of this. The software guys will laugh and say, "give me a 3-axis rate gyro and accelerometer and we can build it with no moving parts."
And I say that as a software guy who has a handful of resolvers in a drawer somewhere :-)
I think people radically underestimate how much capability we have. It's rarely necessary to do things exactly the way they were done 60 years ago.
There is still a certain percentile doing clever things.
I bet a car mechanic in the 60s would have a hard time to replicate this too.
Its always hard to talk about things or people „in general“. Sure, there are some smart people now and there wwre lots of not so smart people back then. This was not a statement that needed to be disproved. This was a statement about engineering culture in general
150 years ago an engineer was a guy with huge moustache and a hat responsible for 200 workers building a bridge that should last for 200 years.
Times change, but there is not much sense in comparing.
The strong impression I always get from the entire Apollo program is "they didn't know it couldn't be done at the level of technology available, so they did it anyway".
That and they essentially had unlimited money.
And some guys who had previously succeeded at solving difficult problems under time pressure with limited money. Well, the money was more limited on one side of the war.
There's a nice lecture from Dan Gelbart that discusses things that people thought were impossible, until they were invented:
https://www.youtube.com/watch?v=7ZeBWJLRXqM Creative solutions to impossible engineering problems
I'm willing to bet that these days, that transformation would be done in software! If you have to drive the ball, you'll still need the motors, and the resolvers for feedback, but the coordinate system transformation is "just math."
This really isn't that different than many software tasks, just a different set of basic tools and jargon. When non computer people read a Haskell article that's how it sounds to them.
1960s technology, designed and made in the USA. It seems that people back then were far more clever at making do with what they had.