See list attachedMay 15, 196969-PA-T-78APA/Chief, Apollo Data Priority CoordinationSome “improvements” in the Descent preparation procedures
As we wade deeper and deeper into Descent Mission Techniques, one thing coming into focus is that, of all IMU error sources, the two that hurt the most are accelerometer bias and y-axis (pitch) misalignment at PDI. Having recognized this, we are now proposing some specific procedures to minimize them. This memo is to tell you all about it in some length, I'm afraid.
There is no better test bed for determining accelerometer bias than a spacecraft in orbit. Any output from an accelerometer is bias and procedures have been well established for monitoring, selecting, and updating the accelerometer bias compensation terms in the LGC. On flights prior to G, the practice has been to establish a threshold below which the compensation would be left alone and above which it would be updated from the MCC. Many of us now feel, and I am proposing that on the G mission, it should be standard procedure prior to DOI for the MCC to update accelerometer bias compensation terms in the LGC routinely, regardless of how good or bad the currently stored values are. The threshold is zero.
Pitch misalignment is a little bit tougher. May I first just state some facts to build on?
a. The current Mission Techniques provide only a coarse IMU drift check by comparison of the docked IMU alignment at DOI – 2½ hours to the undocked AOT alignment performed at DOI – ½ hour. The docked align- ment uses the CSM IMU as its reference and has an estimated accuracy of 0.5° in all axes, so drift rates as large as 0.5°/hr could go undetected. (Specifically, the accuracy of this drift estimate is ± .25°/hr.) PDI occurs about l½ hours after the AOT alignment, which means it is possi- ble for pitch misalignments like 3/4° to build up. That's sort of a worst case kind of number, and to quote such a value will drive statis- tically minded people out of their gourds, but it helps me make a point.
b. Tolerable pitch misalignment at PDI to support a successful landing is in the order of 1° assuming the landing radar comes in early enough to compensate for the dispersions that have built up.
c. Descent aborts become hazardous if the pitch misalignment at PDI exceeds about 0.35°. (This number is being more accurately deter- mined, but I'll bet it comes out within 0.05° of that guess.) This is assuming the worst abort situation, namely aborting at an altitude of about 13,000 feet because no landing radar data has been accepted. If we are willing to go beyond that point with no landing radar, the tol- erable misalignment is smaller than that. The point is that the IMU performance requirement to support descent aborts appears to be the more constraining than to support descent itself and I think we all feel that it is intolerable to continue descent beyond the point a safe abort could be executed with the degraded PGNCS.
d. Since the AGS has to be aligned to the PGNCS prior to PDI, and pitch misalignment in the PGNCS has an equal effect on the AGS. They are not independent in this respect.
e. Given high bit rate telemetry, ground monitoring techniques are adequate to detect an unacceptable IMU misalignment within the first two minutes of powered descent. Thus, the crew could be informed and instructed to abort safely.
f. To abort a lunar landing mission, if it could have been saved by improving procedures, is rather unacceptable.
Based on all that, we have two recommendations, either or both of which should help the situation considerably.
The first is a proposal for a better docked PGNCS alignment suggested by Bob White of MIT, which should allow us not only to detect a drift- ing IMU, but to update its compensation such that we may proceed with a nominal mission. Detailed procedures development and performance analysis is under way at this time. It will demand some modification in the crew timeline during the LM activation and checkout period as well as the implementation of a new RTCC and/or ACR computer program and MCC procedures. The technique requires two spacecraft attitude maneuvers while in the docked configuration with the LM and CSM crew simultaneously keying out CDU angles before and after each of these attitude changes. All of this must be done after the LM IMU has been coarsely aligned as in the current flight plan. With this data, the flight controllers can compute the LM IMU orientation and torquing angles required. This technique is expected to be as good as an AOT alignment. It does not require knowing the relative orientation of the two navigation bases nor reading the docking ring index!
The other proposal involves making a drift check prior to PDI; it requires no MCC participation. Considerable effort was given to including an IMU alignment in the timeline but many of us have concluded the lighting conditions make it chancey at best. The only place it fits in the timeline is from PDI – 30 to PDI – 15. This period is almost perfectly centered around local high noon. Either the sun or the moon is in the AOT field of view for almost this entire time, making use of stars almost impossible. Except the sun! The nice thing about the sun is that it is certainly visible. Also since the whole mission profile is keyed to lighting regardless to landing site and month of the year, the sun will always be located in the same place with respect to the LM. MIT has been asked to write up a precise step by step procedure for doing this. Essentially it consists of the following:
After entering the descent program (P63), the crew would accept the option offered them to go into the alignment program (P52). They would specify the sun as their first “star”. The LGC has the solar ephemeris and will control the spacecraft attitude to place the sun in the center of the AOT. (The rear detent position should probably be used to minimize attitude change unless we do PDI with windows up.) The crew would readout the CDU gimbal angles to which the LGC is posi- tioning the spacecraft; of particular interest is DSKY register No. 2 – the y-axis. The crew would then take over attitude control and cause the sun to cross the AOT retical line in the pitch direction at which time the actual spacecraft CDU angles would be keyed out on the DSKY. The difference between this actual pitch CDU angle and the previously noted predicted value is a direct indication of drift since the AOT alignment one hour earlier. The mission rule would be: if indicated misalignment is less than 0.25°, the nominal mission should be con- tinued; if the indicated misalignment exceeds that value, PDI must be delayed one rev, an AOT alignment would be performed two hours after the previous one and the MCC would determine and update the PGNCS drift compensation prior to LOS.
The value of the first recommendation is that it provides a chance to detect and fix a problem without perturbing the nominal mission. The value of the second is that it allows detecting and fixing a problem before PDI is attempted, although in the worse case it forces delay of PDI one rev, which I am sure we are going to find is a highly undesirable thing to do.
That in a million words-or-less is where we stand on this matter today. We will continue our analysis and procedures development based on this. One unfortunate fact is that if we adopt these proposals, they will not have been tested on the F mission, but I think we would all be naive if we thought we are not going to learn things on F that force us to change the procedures anyway.