Chapter Nineteen: Apollo XIII
After the Countdown Demonstration test (CDDT), which started on March 13, 1970, a detanking of the two cryogenic tanks commenced. The first tank dropped normally to fifty percent while the other tank only drained eight percent out – a very small amount. This was attributed to an internal leak path causing refilling of a tank while draining the lox out. There had been an earlier incident in handling the tank, well before installation in the vehicle, and it dropped two inches. This supported the scenario of a mechanically caused misalignment of the concentric tubes in the tank, allowing lox to leak back into the tank and mostly offset the attempt to drain lox out.
So now, for the first time, the heaters in the tank were used to pressurize and expel the lox. Back in 1966, the voltage for the heater was increased from twenty-eight volts to sixty-five volts in order to aid in faster pressurization and lox removal. The sixty-five volts were supplied via a “ground-test-only” harness. However, the sixty-five-volt GSE harness also powered two thermostatic switch circuits, whose function was to protect the tank temperature from exceeding eighty degrees Fahrenheit by opening and unpowering the heater circuit. However, the thermostatic switches were never modified, qualified or acceptance tested at sixty-five volts. It was a serious error in our system of making changes and a reminder of how seemingly minor changes can propagate into something much worse than the situation you think you are improving,
As the tank was drained with the heaters on, the tank warmed and the thermostatic switches tried to open but were welded closed by the sixty-five-volt acing. Therefore, the heaters continued to draw power for hours and the teflon insulation protecting the fan motor wires inside the tank was mostly melted and destroyed.
A partial fill test was conducted on March thirtieth with the same signature of slow drain of the second tank. It was concluded that the leak was internal to the tank –partially refilling instead of draining – and we would not be draining the tank during flight. Therefore, there was not sufficient reason to replace the tank.
And so, the stage was set for the failure to come during Apollo XIII.
It also became more apparent that the media coverage and perhaps the public interest had cooled noticeably from the run up to Apollo XI. The timing was ironic because after some long period of deliberation – maybe years – NASA had agreed to allow two journalists to sit in a small booth within the viewing room, overlooking the floor of MCC. This was never much of an issue for us because the media had full time access to the air/ground loop with astronauts/capcom traffic and the Flight Director loop. So the granting of access added a real visual of the room, available to two journalists and, presumably, more personal than the ever-present TV coverage of the operations floor of MCC.
Certainly, to me, and I would say most others, it simply did not matter. We were familiar with some regular visitor traffic through the viewing room and we all just ignored it. The journalists’ presence made no difference to us. However, it was disappointing to some that the intensity of coverage diminished after Apollo XI. I thought it was a somewhat natural reaction by the media and it did not bother me. The intensity of Apollo XI media coverage could not be maintained indefinitely.
The Apollo XIII crew was Jim Lovell, Fred Haise and Ken Mattingly through many of our training runs. However, late in the flow, Ken was replaced by the backup Jack Swigert as the CM pilot, because of a medical concern for Ken’s exposure to a child with measles. Jim Lovell was the veteran of two Gemini flights and Apollo VIII. As with all other astronauts, the term “rookie” is really not applicable to the other two crewmen because of their total involvement and training in all the steps leading to this flight and their test flight pilot experience. For example, Fred Haise made it his business to know all about the LM, even to knowing where all the critical wires in the LM were routed behind the close out panels and how to use that knowledge for a hot start if necessary. Jack Swigert was the astronaut office initiator of the malfunction procedure methodology for the CSM. It turned out that Ken never developed measles, but a bias to the cautious side lead to his being bumped from the flight and later assigned to Apollo XVI. Although never done before, the CM pilot was the easiest person to swap out because his critical role at the moon was the solo tending of the CSM while the other two crewman landed. Nevertheless, I am sure it gave Deke and Jim a serious round of discussions.
Milt Windler was the lead Flight Director for XIII and was on duty for the launch phase. The countdown was normal and the Saturn V rumbled off the pad at 2:13 p.m. EST on April 11, 1970. Then, shades of Apollo VI, the second stage center engine shut down more than two minutes early. The Trench was able to verify that the guidance would perform well and the vehicle should burn all the propellant through the other four engines and end up close to a normal orbit, which it did. Then, a “GO for TLI,” the SIVB burned and we had the prospect of a quiet coast out to the moon.
About two-and-one-half days into the flight, I came out to MCC at about 8:30 p.m. CST on the evening of April 13, 1970, expecting a quiet night on the console. Gene’s White team was coming to the end of a long day for the crew, finishing with a narrated TV tour of the LM. They were back in the CSM getting ready for a sleep period. After reading the Flight Director log and catching up with Gene, I went on a walk-around through the back rooms to take the pulse of the team. As a routine procedure at that time, the crew was asked to turn on the fans in the cryo tanks to get a uniform mixture in the tanks for the sleep period. The vehicle was two hundred five thousand miles from earth, eighty percent of the way to the moon and just beginning to fall into the influence of the lunar gravity.
And this was the moment when the bare, and now-powered, fan wires contacted a metal surface in the tank, discharged in the oxygen rich environment of the tank and caused an explosion.
55:55 GET (Ground Elapsed Time since liftoff)
The crew report of, “Houston, we’ve had a problem here” changed the narrative from the start of a crew sleep period to something else – uneasy, but still not clear. Somebody turned to me and said, “Glynn, you may want to get back to the front room –NOW.” I did and plugged in at the Flight Director console to hear a confusing array of multiple indications of problems such as, “Main bus B under volt, fuel cell disconnect, O2 tank low pressure.” At first, it was necessary to be careful and rule out the possibility that some electrical/instrumentation problem was creating the appearance of a bad situation.
56:14 GET (0:19 minutes since problem start)
The fact of a really serious condition began to dawn on the team as the crew reported seeing the spacecraft venting particles out the window. (That’s where the O2 is going and why the O2 tank pressure is so low. And that could be associated with the loud bang initially reported by the crew.) We soon realized that this was not a matter of preserving the landing mission, but this was now about saving the crew. Gene’s team struggled to save what they could of the CSM cryo/fuel cell systems for further use and to reconfigure some of the systems so they would operate properly in the face of the electrical system failures. A CSM power down was started at 56:22 GET and reached a level of forty-one amps.
56:25 GET (0:30 minutes since problem start)
EECOM was concluding that this was not an instrumentation problem and two fuel cells were indeed lost. At about this point, the crew became involved in trying to control some unexpected vehicle rates, which were assumed to be due to the venting.
56:31 GET (0:36 minutes since problem start)
The pressure in the other oxygen tank, O2#1, was reported low and still dropping. More power down was needed. MCC had the crew turn on tank heaters and then the fans to try to arrest the pressure loss – but to no avail. Minutes later, the CM O2 surge tank was isolated to conserve it for entry. We had only one fuel cell and its supply tank of cryogenic oxygen was expected to go to zero in two hours or less. It was near time to start using the LM as a lifeboat. But a few things remained to be done first.
In t
rying to find a way to assist Gene and his team, I was already engaged with Jerry Bostick who was sorting options with the Trench for how to return home from this point. Jerry guided the Trench team through the options. John Llewellyn was also on scene to ride shotgun with Tom Weichel. John was able to focus on the downstream decisions while Tom was occupied with the immediate aftermath of the problem. It is very easy to understand that there was a very strong sentiment in MCC not to go to the moon, but to turn around, and get on the way home ASAP.
Understandable as that attitude was, it would take about six thousand feet per second to perform the necessarily very large maneuver. The only propulsion system with that much power was the Service Propulsion System (SPS) located in the service module. And we had some real concern that the service module had been damaged in whatever had caused the original loud bang. But more importantly, there was a limited amount of power in the CSM entry batteries that would have to be used for a powered-up SPS propulsion maneuver, about fifty amps. A major burn is normally done with the higher power capability of 1 or more fuel cells, but the last fuel cell was fading fast. The necessary electrical power drain would probably come close to depleting the small entry batteries (the only power available for entry) and we did not yet know if they could be recharged. And as another decisive negative consideration, in order to make the burn achieve six thousand feet per second, it would be necessary to jettison the mass of the LM descent stage, which contained most of the batteries and cooling water needed for the trip home. I summarized this situation for Gene, as described above, with Jerry’s help and the Trench confirming the situation and our assessment of options. This was not even a close call. We had to go around the moon.
56:48 GET (0:53 minutes since problem start)
Gene agreed and announced the go around the moon decision to the team.
57:05 GET (1 hour and 10 minutes since problem start)
With full recognition of how demanding this situation was, The Black team and I came on duty. Positions were manned by Jack Lousma at Capcom, Larry Keyser at AFD, Gary Scott and Ed Fendell at INCO, Bill Boone and Maurice Kennedy at FIDO, Tom Weichel at Retro, Gary Renick and Will Presley at Guidance, Merlin Merritt at Telmu, Hal Loden at LM Control, Clint Burton at EECOM, Jack Kamman at GNC and Spencer Gardner and Elvin Pippert at FAO and all the other off-shift flight controllers and astronauts who gathered to help within less than an hour of the problem. (Consoles had four jacks for headsets to plug in and they were all occupied.) This flight control team was a solid set of operators, but hardly any had been in this kind of circumstance.
Chris and Sig were also there by this time. Besides the flight controllers, we also had the best brains available through our offline SPAN communications and data network with all of the engineering and program organizations, both in NASA and industry, and from all regions of the country.
The MCC was full. Even so, the comm loop discipline was good. No illusions by now – we all knew that this was a very big hill to climb. And it was ours to handle. It was time to get on with it. The situation was:
A loud bang was reported at the start of this problem and eventually the crew reported.
Particles were venting from the SM.
O2 tank #2 was at 0 pressure. O2 tank #1 was predicted to last no more than two hours.
Fuel cells 1 and 3 were not supplying power.
Main bus B and AC bus 2 were zero, since both were fed by fuel cell 3.
Considerable reconfiguration had been performed to get enough thrusters on main Bus A.
The trajectory was not on a free return to the entry corridor and it needed correction.
To help in understanding our response, actions can be considered in two categories:
Managing the Configuration of the Spacecraft Systems – Sometimes the configuration choices were driven by troubleshooting of problems (e.g. leak isolation, switching redundant paths, preserving capability, etc.) – sometimes to support a mission need (e.g. a propulsion maneuver, a power down, the optimum control system capability, etc.). The mission need to operate within the reduced consumables also stressed the configuration choices well beyond the normal. The choices also required closer coordination among the flight controller positions in the MCC because the window for an integrated solution (balancing the demands of propulsion, electrical power, time, coolant water, guidance equipment, communications, etc.) was much narrower.
Return Home Steps – The necessary mission steps to return the vehicle and crew to a safe landing.
In the minute-by-minute voice traffic within MCC and with the crew, the return-home steps provided an overall mission framework towards the primary goal. But, the majority of time and interactions was spent evaluating, deciding and implementing the spacecraft configuration choices to stabilize and/or improve our posture to support the crew and accomplish the return steps. And burned into us from years of training and operations was the CARDINAL rule – “Don’t screw anything up and make the situation worse than it already is.” It also helps to understand the evolution of the return-home plan as an incremental process. We did not begin by having a comprehensive plan – but, rather, took steps as we judged them to be necessary, appropriate or keeping us with the best range of forward options. Think of the process as the fog clearing enough to commit to the next step. For example, the decision at 56:48 GET to go around the moon, rather than attempting a direct return, was the first of the RETURN-HOME steps. This STEP #1 was driven by a fact-based analysis of options illustrating that the go-around option was the only workable one. Some of the next steps were more based on the judgment (without a full factual analysis) of what was best. I will highlight each of the return-home steps as our shift and the flight progressed, with the first one being Gene’s earlier decision to swing around the moon.
With the vehicle rates under control with a new RCS configuration, the Black team first focused on the last-ditch steps to try to save some of the CSM cryo-fuel cell capability. The last step was to close the reactant valves to the non-performing two fuel cells in an attempt to isolate the possibility of an O2 leak inside the fuel cell itself. Once closed, the fuel cell was without the fuel to run, and could not be restarted. No joy on the first cell and then the second one ended up with the same negative result. The O2 leak continued.
57:35 GET (1 hour and 40 minutes since problem start)
The two LM crewmembers were making their entry into the LM at the same time we were calling up that recommendation. “We’re already on our way” was the reply. At this time, we had one good fuel cell #2, but the oxygen pressure to feed it was still dropping. The crew began the initial activation of the LM, designed to get the batteries, life support systems, and communication/instrumentation systems online. MCC received the initial LM telemetry signal at 57:57 GET. On a personal note, in managing and prioritizing the flow of comm traffic with the crew, Jack Lousma was a pillar of stability for this team over the course of a very long night.
During the power-up, I had a short time to consider options. My first strong inclination was to power down quickly, conserve LM consumables and work out a plan. But, I also had serious concern that the venting particles would preclude getting a good guidance platform alignment for the burns that would be required later. Tom Stafford was intense about getting and maintaining the alignment in the LM. There were confirming nods from Jack Lousma and other Apollo crewmen. This was now our only opportunity to get the CSM inertial guidance alignment transferred to the LM guidance system, even if we later decided that we did not want to use it and powered the LM platform down. And I decided to take the time and electrical power from both vehicles to accomplish that transfer and then decide what was next. If not done now, this opportunity would be lost and no longer available. It was much too early to foreclose this option going forward.
57:54 GET (1 hour and 59 minutes since problem start)
The CMP powered down as much in the CSM as he could while keeping the CSM guidance system up. Because of the decreasing O2, Clint
Burton at EECOM was watching to see a degradation in fuel cell 2 in order to know when to put an entry battery on to support the electrical bus. We intended to stay “up” in the CSM until a LM guidance alignment was transferred. Once we were on entry battery A, we wanted to minimize the number of amp hours withdrawn from it since we did not know if we could charge it from the LM for later use. The alignment transfer itself was a tedious process of crew/MCC coordination and the reading and checking of a lot of numbers as they were entered into the LM guidance computer. During the period of transferring the alignment, there was a short period with neither of the attitude control systems on. This was quickly recognized and corrected. It was upsetting that we (and I in particular) missed that condition. Although no real harm attended, we could not afford mistakes.
Around the time of these final CSM closeout steps, I had a brief period when the severity of the problem really struck home. For the first and only time in ten years of console experiences in training and actual flights, I had the sense of the bottom falling out from under me and my stomach heading for that dark hole. I would like to believe that it was due to an acute awareness of the “Abandon Ship” situation. But the feeling was emotional, not intellectual. “Holy xxxx. I can’t believe this is really happening.” Scary – but the ten years of experience kicked in and it took about ten to twenty seconds to return from that place.
58:40 GET (2 hours and 45 minutes since problem start)
Jack Lousma helped get the LM team back on tailoring an existing checklist for LM power-up and it was now time to turn the CSM power off. We had used about twenty amp hours – or fifteen percent of total entry battery power – before power down. The CM was going to get cold and uncomfortable and we still needed Odyssey to get home.
Highways Into Space: A first-hand account of the beginnings of the human space program Page 19