Early planning for Apollo: how could they think it would work?

Factual Questions
1

One Giant Leap (by Charles Fishman; it’s at least tied for the best book ever written) tells how John Houbolt had to struggle to get Apollo planners to choose Lunar Orbit Rendezvous. Roughly everyone else was debating the other two possibilities: one rocket all the way to the Moon and back, or one rocket all the way from Earth orbit to the Moon. In other words, everyone else figured the astronauts could land a big rocket on the moon, then take off in that rocket and return to Earth.

How could that possibly happen? The rocket had to land on its tail, didn’t it? And the astronaut(s) had to be somewhere near the top of the rocket? Even if they could find a level-enough landing site, quickly enough, how would they keep the rocket from toppling over on touchdown? If they managed that, how would they get themselves down to the surface? If they managed that, and climbed back up, is the rocket somehow sure to be vertical? If not, how do they steer it to get it pointed the right way after takeoff? How do they know the right way?

All obvious questions, that must have occurred to them. What were their answers?

2

All kinds of weird and unworkable ideas were floating around the space program (edit: no pun intended!). But they weren’t all known to be unworkable right away.

I think Houbolt’s story is more instructive of bureaucracy and how good ideas often have trouble rising to the top. He made a Hail Mary pass and it actually worked. That’s often another story in many organizations.

One could also make the case that we did Apollo the wrong way, meaning not in a way that was sustainable. It was a crash program and they did what it took to make it happen quickly, not optimally. As much as I’m a huge Apollo fan, I wish it had been done differently, without so many throwaway components.

3

Their answer was to decide not to do it that way? Because maybe they realized it was too difficult?

BTW, what you described is pretty much the way they plan to do it the next time, using an Elmo Silo.

4

You don’t need to land a big rocket on the Moon. You can do it with a small rocket. Basically, think of something just a little larger than the lunar lander we actually used, except that the crew would stay inside of it for the whole flight, not just the part on the surface.

5

I believe the Starship HLS is as large as it is because the intent is to carry large amounts of supplies for a lengthy stay. The crew capacity is planned as only 2 to 4 astronauts, but future versions may have expanded crew capacity, trading it off for cargo already delivered.

Or, the whole thing may be scrapped. This is Elon Musk, so who knows. According to his vision, he has a colony on Mars already.

But some version of Starship HLS is currently endorsed by NASA.

6

There were a lot of options considered for the lunar landing. My favourite is the opposite direction. The LM was essentially a rocket with handlebars. Astronauts did the entire mission in their spacesuits.

Designs for one or two suited astronauts standing on a platform holding on were around.

This paper from 1986 does a nice job revisiting the choices made with more modern systems engineering tools. LOR still wins. Interestingly the least risky choice was direct ascent with the mooted by never developed Nova rocket. It was considered that any mode would work given enough money and time.

A big driver for choices was that the progress on much of the project had already let contracts and made design choices. The Saturn 5 was in development and the CSM had already begun development. Changes here would have delayed the project past the end of the decade. So LOR was the best and least risky option with what they had. EOR was the worst because of the additional risk of untested technology and techniques. In orbit fuel transfer being a big one. Which it still is. 50 years later and still not demonstrated. But if it can be made reliable it opens up other modalities.

7

You raise some good questions addressed by others above. But this bit is not relevant.

All those problems have to be solved no matter what sort of lunar lander they had.

Rockets only want to swap ends and crash. It’s only active guidance and stabilization that keep it going the right direction. And that’s only possible when you have a way to determine what that right direction is.

8

Yeah, controlling rockets is counterintuitive, but not hard.

I’m reminded of a story from R V Jones’ book Most Secret War, where in the UK he and a team were looking at intelligence reports about the V2 rocket. Jones calculated that in order to hit anything in the UK the precision needed to align the rocket for takeoff was essentially infeasible. Needing seconds of arc precision to loft the rocket on a ballistic trajectory. Then someone else in the team muttered “gyroscopes” and Jones tells of how suddenly everyone realised that the V2 threat was real and all started shouting “gyroscopes!!”

The Apollo LM had exactly the same issues with off vertical landing attitude as any of the other design. It is just that a leaning over LM doesn’t look very dramatic because it is so squat. But 10 degrees of lean is 10 degrees of lean, and your guidance system has to be able to manage it on liftoff no matter how tall your vehicle is.

Luckily for Apollo the basic technology for guidance was already well understood, thanks to the ICBM projects. The Apollo guidance system was more advanced than an ICBM in every respect, but this was a matter of scale, not intrinsic technology. In order to navigate in 3D space the guidance system needed an additional gyro axis (and arguably could have usefully had a fourth to avoid gimbal lock, but that was a nice to haver and would have added significant size.)

The Apollo guidance computer was a minor miracle of miniaturisation and genius level electronic design. Not just the processor, but the peripherals and ancillary equipment. The CSM and LM had identical AGCs and gyros and acelleromter systems. Moreover the LM had a standalone guidance computer that had one job only, get the LM back into orbit where they could be rescued by the CSM. The LM could track the CSM in orbit and control the ascent to match.

There was, in addition, the Saturn V’s guidance system that resided in a ring atop the booster just between the LM and the rest of the Saturn. Famously, when the Apollo 12 launch was struck by lightning, dropping the entire CSM electrical system off line, including the guidance computer, the Saturn just chugged away and lofted the payload exactly as needed. That sustem was different again. Built by IBM it was much larger and heavier, and very simple - not really being a fully fledged computer system. But robust.

Any issue with guidance and managing launch from the lunar surface was a solved problem no matter what the actual vehicle used was. A big tall rocket looks as if it should be much more precarious, but assuming it doesn’t simply fall over on landing (something some modern unmanned missions have had problems with of late) launching form the surface isn’t an issue.

Controlling a rocket is the same problem as balancing a broomstick on your hand. A bit of practice and you can tune your reflexes to make it look trivial. An electro-mechanical version of this is a staple of undergraduate control theory labs. Students get to code up the control law, work out the parameters, and get the machine to swing the stick upright and keep it there, even when you push on the stick. Very satisfying.

This is also why rocket motors are on the bottom of rockets. Counter-intuitively, this is where they must be for control to work. Intuition would suggest that putting the rockets at the top and pulling the craft would be more stable. But it isn’t. The system will develop worse and worse oscillations as the pendulum effect of the vehicle hanging down dominates.

9

That’s very interesting. I’ve been reading about rocketry for a long time and never considered it.

The balancing a stick on your hand analogy is good too.

10

Also, you can make even a very tall rocket stable by putting on long legs. This isn’t usually done on Earth for aerodynamic reasons, but that’s a nonissue on the Moon.

11

Possible solutions for all of these problems were depicted on the covers of Science Fiction pulp magazines for decades before NASA started planning a trip to the moon. Retro rockets would keep the craft vertical. Extending legs would emerge from the craft before the rockets were turned off. An elevator or bosun’s chair would lower the astronauts to the surface. Before takeoff, the extended legs would orient the craft in the right direction. Or a hundred other solutions thought up by the those writers and artists. Some of those other predictions included things quite similar to the lunar lander.

12

I apologize if this is off-topic, but one question I’ve had for decades is what coordinates the guidance computers used; earth-centered, moon-centered, rectangular, polar, or something else? And that’s just for position; what did they use for velocity and orientation?

13

That tickles me.

14

The Apollo Primary Guidance, Navigation, and Control System (PGNCS) and Apollo Guidance Computer (AGC) used two primary coordinate systems:

  1. Basic Reference Coordinate (BRC) System: An orthogonal, Earth-centered or Moon-centered (depending on mission phase) inertial coordinate system.

  2. Stable Member Coordinate System: A localized, three-axis Cartesian coordinate system tied to the Inertial Measurement Unit’s (IMU) gyroscpes.

To correct for gyroscope drift, sextants were used for celestial navigation.

Those gyroscopes and celestial navigation units also take care of orientation. Velocity was determined using gyro accelerometers. The AGC kept track of all acceleration events from firing thrusters/engines and integrated them over time to determine instantaneous velocity and trajectory vectors.

15

And if you’ve played Kerbal Space Program you may know how fiendishly difficult that is.
I could never do it manually, had to download a mod that includes an autopilot “mechjeb”.

16

I don’t understand how anyone ever did anything without the Mechanical Jebidiah.

17

I managed to land on the mun and get back to Kerbal, but with the same stage landing in the mun and going back.
To implement LOR I HAD to use our friendly mechanical kerbin.
And I think landing in Duna and going back withouth Duna Orbit Rendevous would’ve been impossible, so I practiced with mechjeb until I got it working in the Mun and then sent the mission to Duna.