Speaking of the "chain of human choices", this snip from Gene Kranz's book about the lead-up to Apollo 11 caught my eye last night:
Flight Readiness Review was conducted on June 17, and there were no major open items. The review went well until Kraft made a few comments about the landing data rules. A free-for-all started, and I was called on to write some specific rules on the communications and data requirements for landing. This issue continued to be debated until the week before flight, and it appeared that some of the folks at headquarters were getting damn nervous about the consequences of a crash, if one occurred.
Chris, Cliff, and I agreed on the real rule: that “we must have enough data to reconstruct what went wrong.” This rule left me the maneuvering room to take it right down to the surface before I had to make a land or abort call. Once we were close, I intended to let the crew go if everything appeared okay to them. I considered a low-altitude fire-in-the hole abort more risky than landing without data. I always looked at a fire-in-the-hole abort the same way I looked at a parachute when I was flying jets. You use a parachute only when you have run out of options.
The day before the launch, I processed a write-in mission rule change that legitimized this landing philosophy: “The flight director will determine if sufficient data exists to continue the landing.” No computer could make this call—it had to be a human decision. [Emph. added]
I'm just saying the same things that need to be said again and again with fierce conviction. But this isn't a post about human decisions and accountability. It actually got me looking into the data requirements, how NASA developed mission rules, and whatnot.
I was curious if there were any more details available about the debates, and Kranz himself filled in some gaps on CSPAN back in 1999 (I've clipped the interview, and kinda sorta tidied up the transcript):
[H]eadquarters people, the program managers, as well as Chris Craft, was concerned that if we would crash and not have enough data to figure out why we crashed, we'd be in jeopardy of...not only losing the lunar goal, maybe the entire program. So, everybody wanted to make sure that there was some formula that would be used by the team to say, okay, we got enough day data to continue.
I fought this particular rule, because they wanted...some numbers with this thing and I fought this rule all the way through the process of building the rule...going through mission reviews, et cetera. I wanted a simple one that said, the flight director will determine whether sufficient data exists to continue the mission...I just wanted...it that simple, that it was a subjective call by the flight director. and this was batted back and forth until very close to the mission and it was not resolved, so, I wrote into the mission rules that exact statement, the flight director will determine if sufficient data exists to continue.
Well, going back to the landing day now, this adequate information means voice information and telemetry. as soon as the spacecraft cracked the hill and we were now silently coasting down to the 50,000-foot mark above the moon, the telemetry was broken, the voice was broken. We couldn't communicate. It seemed nothing was going right.
And immediately that rule came to mind, do I have sufficient information to continue, but then we'd get a bit and I'd say, ah-ha, we can look at the spacecraft and there were a couple of times I would make calls, okay, all flight controllers, go-no go time, use the last valid data points that you saw. Well, this might be 30 seconds old. So, they're making decisions based on stale data.
We kept working, trying to figure out what was the problem with the communications. and this turned out to be bad information on the attitudes used in the spacecraft, because we were getting some [radar] reflections off the skin of the lunar module. But again, this is too late, we had to try to solve the problem in real time and again, go back to the team work.
Charlie Duke, who was my spacecraft communicator, was looking at the signal strengths and he saw the signal strengths vary - he had also worked the Apollo 10 mission - he suggested to Don Puddy - who...had the responsibility for the communications, but also the life support, electrical system, on the lunar module - do you think we could...make an attitude change, would that help any? So, then we tried an attitude change.
Fortunately, in training, we had also worked in relaying voice information from the ground to Mike Collins back down to the lunar module. So, we were using every conceivable way to communicate. In the meantime, time is marching down to my go-no go points.
The whole purpose of data is to provide humans with actionable insight1, allowing us to make wise decisions, often during crisis. So it was a good thing they thought ahead of time about all the data requirements and what the controllers could do in myriad cases, especially during that singular shot at landing (I've got another post more about that percolating in my noodle right now).
Kranz and team spent a year(!) just figuring out all the stuff like what data was mandatory and what was "highly desirable" to know so they could determine the position of the LM on the lunar surface. They also needed to understand what impact changes could have on the mission:
To design a flight plan that is operationally usable for performing a mission, operational data are required for the various systems. Mission objectives require various systems to be used in varying modes; therefore, an understanding of how the systems can be used is required to determine if the objective is feasible and compatible and could be scheduled in the flight plan...
On the Apollo 12 mission, a new program was added to the onboard computer that provided the capability to program a given spacecraft pitch rate automatically. However, in examining the operation of this new capability, a systems anomaly was found. The anomaly was that the automatic-pitch-rate program was economical, with respect to propellant, only in the heads-up orientation (0" roll) and was not suitable for use heads down (180" roll).
Thus, the attitudes in which this new capability could be used were constrained by these operational data. The operational performance data were also useful in pointing out that this pitch-rate capability could be modified so as to be useful in attitude orientations other than 0" roll without prohibitive propellant consumption.
The mind boggles at the complexity, and how NASA was able to work through all the problems, one by one, until they succeeded. In conclusion: if they can put a man on the moon, then why can't they2...
1 - Sometimes even our metro bus service has problems with that (I did the math: 29,423,118 minutes = 55.98 years).
2 - I had that issue of Mad Magazine, one of a handful I collected as a wee lad.
No comments:
Post a Comment