The Digital Autopilot

Shea made one other change in the control system for the Block II design. When he came to Houston from headquarters, the Apollo guidance computer was just that—a computer that figured out where the spacecraft was and where it ought to go. Another system actually took care of commanding the thrusters to move the vehicle and was under development at Minneapolis-Honeywell, the same company that had made the control systems for Mercury and Gemini. As in those earlier spacecraft, if the astronaut wanted to orient the spacecraft, he would command the Honeywell servos to go to a particular point, and then analog feedback loops would command the thrusters to hold it there. As in Gemini, an indicator would show velocity to be gained to the astronaut, who would command thrust while the indicator counted down toward zero when the maneuver was complete, a scenario appealing to the pilots. But the Honeywell project was behind schedule and having technical problems, and NASA engineers began to think about simplifying the situation.51

The IL system plus the Honeywell system meant that the Apollo spacecraft contained two nonredundant control computers, two sets of gyros, two sets of electronics. Why? Cline Frasier was the NASA representative on the project: ''I was always nervous about all of these mechanical parts running around.'' His first idea was to put the MIT and the Honeywell computers in parallel rather than in series, so that if the analog one failed the digital one could still fly the craft.52 But Frasier went one step further and convinced Shea to eliminate the Honeywell autopilot and incorporate its function into the digital computer. Shea had been the program manager for the Titan missile, which used digital computers for flight control, so he was no skeptic about their capabilities.

ASPO manager Clifford Duncan made the digital autopilot decision in June 1964. ''So Cliff [Duncan] sent me to the Instrumentation Lab to go up and tell the guys,'' Frasier recalled, ''that instead of two computers there is only going to be one. Instead of just doing the guidance, they were also going to do digital autopilot.''53 Though not eliminated, the Honeywell SCS was much reduced in complexity to provide a simple analog backup to the digital autopilot.54 For the IL engineers, this change amounted to doubling the responsibilities of their computer, and hence doubling the work to be done before the flight (fortunately, NASA ordered the memory doubled as well, but the IL still had to write and test the programs).

The digital autopilot made the IL computer central to the Apollo spacecraft—some compensation, perhaps, for no longer being the primary source for navigation. Now, the controls of the CSM and the LM would go through the AGC and its software. A software-based ''state estimator'' could derive the motions of the vehicle and automatically compensate for thruster failures and other anomalies. Shea cemented the decision in a memo at the end of 1964. From now on, the ''digital autopilot control mode,'' would be primary, with the analog Honeywell system as a backup.55

Putting a digital autopilot into Apollo was a radical step. Today, some commercial airliners use this fly-by-wire technique, but in the 1960s, no aircraft had yet demonstrated digital fly by wire, and only a few research aircraft had used an analog form. The astronauts were none too happy. ''Why don't you guys quit wasting time,'' David Scott recalled thinking about the decision, ''go back to MIT and think.''56 Frasier believes the digital autopilot never would have been approved were the flight crews not too busy with Gemini to notice. He remembers running into Pete Conrad one day in a hallway, and Conrad ''just chewed me up one side and down the other and said that we're crazy. That his friends at Grumman had told him this stuff wasn't going to work, and it was his life on the line, and kind of on and on.'' Frasier, a junior engineer at the time, was genuinely intimidated at being told off by an astronaut, but the decision had been made.57 The astronauts could make their peace with it, or they could choose not to fly—which none of them did.

As Davey Hoag put it, ''The way we look at it, we actually have four autopilots.''58 The CSM and LM each had two separate modes: free fall mode, which provided attitude control while the vehicle was coasting or in orbit, and thrusting mode, which controlled the large engine on the command service module for major maneuvers. Hoag might have added two more for the CSM: a boost mode, where the AGC monitored the Saturn rocket's performance, and even allowed an astronaut to manually takeover if the Saturn's guidance failed in the second or third stage (a vestige of the piloted-boost debate), and a reentry mode, which worked similarly to the Gemini computer. For any particular mission mode, the navigator would select the program using the computer's keypad.

In Gemini, the computer calculated velocity to be gained and the astronaut controlled thrust with his stick, according to the display. In Apollo, the digital autopilot oriented the spacecraft automatically and then commanded a program to fire the main engine. The digital autopilot also trimmed the large engine's gimbals on a regular basis while it fired, ensuring the thrust was aligned with the vehicle's center of gravity. During coasting flight, the computer could hold a particular attitude, or change it constantly at a defined rate. The digital autopilot also had ''jet selection logic'' in the software to determine which of the sixteen thrusters on the spacecraft should fire at any given moment. If any of the thrusters failed, or if an entire ''quad'' cluster failed, the autopilot would automatically sense the situation, compensate accordingly, and avoid doing anything dangerous. It limited the rate of motion in any axis, to avoid tumbling, and it could automatically recover from a tumble by slowing the rates in each axis. Like the X-15 adaptive autopilot, the digital autopilot contained a model of the spacecraft's motion and compared its actual motions against the ideal in the model, taking into account the effects of body bending and fuel slosh in the vehicles.59

MAJOR UNITS OF TX CM

MAJOR UNITS OF TX CM

Figure 6.2

Physical units of the Block II guidance and navigation system in the command module. (Draper Laboratories/MIT Museum.)

Figure 6.2

Physical units of the Block II guidance and navigation system in the command module. (Draper Laboratories/MIT Museum.)

This new role for the computer had great impact on the IL engineers. Suddenly they were responsible not only for alignments with the stars, but also for detailed interactions with the other components in the system. Suddenly they needed reams of highly accurate data—on the vehicles' dynamics, bending modes, actuator performance, engine gimbals—creating a whole new level of complexity in their relationships to North American and Grumman. About 10 to 30 percent of computer capacity would be required to run the autopilots, capacity that was far from plentiful as the software edged close to its performance margins (figures 6.2 and 6.3).

The digital autopilot also confirmed the decision to use a general-purpose computer in the first place and underscored the intimate links between systems engineering and digital computing. Engineers could move particular functions out of hardware devices and into computer programs, saving critical quantities of weight, money, and hardware complexity. In one example, Shea nixed an expensive program to add a heat shield for the side of the command module facing the sun. With his knowledge of

Figure 6.3

In the Block II system the computer linked a variety of systems, including the astronauts' control of the spacecraft. Note how the manual control signals go through the computer instead of through a separate controller, as in Block I. Not only the computer but also the MIT Instrumentation Lab had to integrate these diverse components, from a variety of contractors. (Hand, ''MIT's Role in Project Apollo, Vol. I,'' 52.)

Figure 6.3

In the Block II system the computer linked a variety of systems, including the astronauts' control of the spacecraft. Note how the manual control signals go through the computer instead of through a separate controller, as in Block I. Not only the computer but also the MIT Instrumentation Lab had to integrate these diverse components, from a variety of contractors. (Hand, ''MIT's Role in Project Apollo, Vol. I,'' 52.)

control systems and the digital autopilot, he simply suggested replacing the insulation with a software routine to keep the spacecraft rotating like a rotisserie, distributing the heat load around the craft. A few lines of computer code replaced a heavy mechanical structure.

In a 1967 interview, Hoag expressed wonder at the flexibility of digital controls: ''All this, and everything else too, in this tremendous computer and we are doing it, by golly, we are doing it.''60 Yet there were costs, mostly in software, in key areas of complexity, reliability, and schedule. All three would raise problems down the line. Still, Aaron Cohen considered the inclusion of the digital autopilot a ''milestone in the Apollo program.'' Frasier believes that without ''the combination of the digital flight control and the navigation from the ground, we would have been at least a couple of years later [getting to the moon].''61

Was this article helpful?

0 0
Telescopes Mastery

Telescopes Mastery

Through this ebook, you are going to learn what you will need to know all about the telescopes that can provide a fun and rewarding hobby for you and your family!

Get My Free Ebook


Post a comment