See list attachedDecember 18, 196969-PA-T-148APA/Chief, Apollo Data Priority CoordinationLM IMU drift checks prior to descent for Apollo 13 and up
We are making some fairly substantial changes to the way we are align- ing and checking the drift of the LM IMU before descent on the Apollo 13 mission. Just for the record I would like to document what it is we are doing and why.
Two things have happened as we progressed from Apollo 12 to 13 which have made it necessary to change the techniques. Probably the most significant is performing the DOI maneuver with the CSM. This in turn presented a problem with regard to landmark tracking by the CSM since we aren't sure it can be done in the 60 x 8 n. mi. pre-descent orbit. So, in order to assure getting the landmark tracking, we decided to recircularize the CSM orbit to the 60-mile altitude. And to get the tracking done in time to use the data, we are forced to undock from the LM 1 rev earlier than we did on Apollo 12. Undocking earlier means that less time is spent while docked during the LM activation and checkout which precludes our making an accurate docked alignment of the LM platform. (We have neither suffi- cient time nor the necessary attitude changes in the new timeline.) On the other hand, undocking early gives us an extra rev of LM free flight which allows carrying out two AOT (P52) alignments during each of the last two darkness passes before PDI.
The other significant thing that made it necessary to change the techniques is the fact that we are landing on a rough area on the moon in which the acceptable touchdown conditions are constrained to a very small area. For planning purposes it is defined as a circle with 1 kilometer radius. The point is, whereas on previous missions we could miss the targeted-landing point by many miles and still land and achieve the primary mission objective, on this flight we cannot even land safely very far from our 1 kilometer circle. This obviously imposes a demand for superior performance from the PGNCS than was needed on previous flights. In particular, we must make sure misalignment of the platform at PDI about the vertical (x) axis is about an order of magnitude smaller than was acceptable on Apollos 11 and 12. On those missions the maximum acceptable x misalignment was based on protecting against continuing with a broken system. Specifically, we were able to tolerate a platform drifting at a rate up to 1.5 O/hr. The fact that this would cause a very large miss in landing point location was not sufficient justification to delay or scrub out the landing. Now we are not willing to go on if the guidance system is going to miss our little circle. An analysis shows that a misalignment in excess of 0.19° at PDI in all we can tolerate. By moving our last platform alignment as late as possible before PDI, we can pinpoint our largest acceptable drift rate. Assuming the latest we can do the P52 is 1 hour and 20 minutes before PDI, the maximum allowable drift rate turns out to be .145 °/hr. (that is about a 4.4 sigma system).
Our number one problem comes about when determining first of all if the system is working better than that, or not. Secondly, if it isn't, how do we get the new compensation to the spacecraft? After a good bit of head scratching, the consensus is that our best determination of drift rate (not absolute inertial alignment) can be made using the CSM platform as a reference while the LM is still docked. If you can assume there is no slipping or bending between the two spacecraft while docked, the MCC is able to detect drift rates in excess of .04 °/hr. dependably. We feel this is at least as good as two P52's spaced 1 rev apart. In fact, it's probably better. So we plan on using the crew's readout (N20) of LM and CSM gimbal angles while docked to make the determination of whether or not the LM IMU is working well enough to support a landing. Furthermore, if we find the drift in excessive, we intend to use that same data to determine new values of drift compensation which will be uplinked to the LM after undocking, but before the first AOT alignment. This procedure should not only be the most accurate way to do it, but also avoids another problem. Namely, there is no straightforward way of using the data obtained from the two AOT alignments, the last of which occurs in back of the moon 1 hour before PDI, and uplinking the new compensation values, if that turned out to be necessary, without delaying the landing 1 rev. Of course we have every intention of rechecking the system for acceptable performance based on the undocked AOT alignments but the procedure outlined above should preclude finding it unacceptable at a time when it is difficult to do anything about it.
We are not changing our criteria used to establish acceptable drift and misalignment about the other two axes, y and z. As before the y limit was chosen to provide a safe descent abort capability since pitch misalignment does not significantly affect landing-point accuracy as long as the land- ing radar data comes in. The z-axis limit is still based on making sure the system is not broken since we can stand massive misalignments around the braking thrust axis. (Note: IMU z is approximately along the LM x-axis.)